2406 09832v1

ON BI-ENRICHED ∞-CATEGORIES
HADRIAN HEINE,
UNIVERSITY OF OSLO, NORWAY,
HADRIAH@MATH.UIO.NO
arXiv:2406.09832v1 [math.CT] 14 Jun 2024
Abstract. We extend Lurie’s definition of enriched ∞-categories to notions of left enriched, right
enriched and bi-enriched ∞-categories, which generalize the concepts of closed left tensored, right
tensored and bitensored ∞-categories and share many desirable features with them. We use bi-
enriched ∞-categories to endow the ∞-category of enriched functors with enrichment that generalizes
both the internal hom of the tensor product of enriched ∞-categories when the latter exists, and
the free cocompletion under colimits and tensors. As an application we prove an end formula
for morphism objects of enriched ∞-categories of enriched functors and compute the monad for
enriched functors. We build our theory closely related to Lurie’s higher algebra: we construct
an enriched ∞-category of enriched presheaves via the enveloping tensored ∞-category, construct
transfer of enrichment via scalar extension of bitensored ∞-categories, and construct enriched Kan-
extensions via operadic Kan extensions. In particular, we develop an independent theory of enriched
∞-categories for Lurie’s model of enriched ∞-categories.
Contents
1. Introduction 2
1.1. Notation and terminology 8
Acknowledgements 8
2. Weak enrichment 8
2.1. Weakly enriched ∞-categories 8
2.2. Tensored ∞-categories 12
2.3. Enriched functors 14
2.4. Enriched adjunctions 17
2.5. Trivial weak enrichment 19
2.6. Tensored envelopes 22
3. Enrichment and pseudo-enrichment 29
3.1. Pseudo-enrichment 29
3.2. Enrichment 31
4. Enriched presheaves via tensored envelopes 36
4.1. Enriched presheaves 40
4.2. On the relationship between weak enrichment, pseudo enrichment and enrichment 42
4.3. Presentability of enriched presheaves 47
4.4. Atomicity of representable enriched presheaves 49
4.5. Transfer of enrichment via scalar extension 55
4.6. Transforming bi-enrichment into left enrichment 61
5. Enriched ∞-categories of enriched functors 68
5.1. Enriched Kan-extensions 71
5.2. Enriched Yoneda-lemma 73
5.3. Enriched Yoneda-embedding and opposite enriched ∞-category 77
5.4. Enriched ∞-categories of enriched functors as an internal hom 83
1
2 HADRIAN HEINE, UNIVERSITY OF OSLO, NORWAY, HADRIAH@MATH.UIO.NO
5.5. Enriched profunctors 87

5.6. An end formula for morphism objects of enriched functor ∞-categories 88
References 91
1. Introduction
Gepner-Haugseng [3] develop a theory of enriched ∞-categories that both at the same time extends
the theory of ∞-categories and the theory of non-commutative higher algebra: the authors prove
that ∞-categories enriched in spaces are equivalent to Segal spaces [3, Theorem 4.4.6.], a model for
∞-categories, and define enriched ∞-categories as a many object version of A∞ -algebras [3, Definition
2.4.5.], a homotopy coherent version of associative algebras, with which enriched ∞-categories share
many properties: any A∞ -algebra in a combinatorial monoidal model category M is equivalent to
an associative algebra in M [16, Theorem 4.1.8.4.]. Similarly, every ∞-category enriched in the
underlying monoidal ∞-category of M is equivalent to a M-enriched category [5, Theorem 1.1.].
By [16, Remark 4.1.8.9.] every presentably monoidal ∞-category V admits a monoidal combinatorial
model. Consequently, as long as one works with ∞-categories enriched in a presentably monoidal
∞-category, which covers most examples of interest, nothing new arises when passing from enriched
category theory to enriched ∞-category theory so far as one is just interested in enriched ∞-categories
but not in enriched functors. On the other hand ∞-categories of enriched functors are generally
tremendously hard to model by their strict counterparts: even in the most basic example of interest,
enrichment in spaces, one finds that the ∞-category of (spaces-enriched) functors is not modeled by
the category of simplicially enriched functors since Bergner’s model structure on simplicial categories
fails to be compatible with the cartesian product [17, Definition A 3.4.9.]. From this perspective
enriched ∞-categories of enriched functors are the real topic and challenge of enriched ∞-category
theory and good control about enriched ∞-categories of enriched functors is highly desirable.
From the conceptual point of view there are several obstructions to the existence of enrichment on
the ∞-category of enriched functors: if V is a presentably braided monoidal ∞-category, there is a
natural candidate for a V-enriched ∞-category of V-enriched functors as the internal hom for a tensor
product of V-enriched ∞-categories: Gepner-Haugseng [3, Corollary 5.6.12.], Haugseng [8] construct
a tensor product for enriched ∞-categories, which exhibits the ∞-category of small ∞-categories
enriched in any presentably En+1 -monoidal ∞-category for n ≥ 1 as closed En -monoidal. On the other
hand if V is a presentably monoidal ∞-category, an important situation in non-commutative algebra
and higher category theory [2], there is no need for any V-enrichment on the ∞-category of V-enriched
functors. However, the ∞-category of V-enriched presheaves on any small V-enriched ∞-category is
always V-enriched as long as V is presentably monoidal. This relies on the fact that V has more
structure than being a V-enriched ∞-category via its closed left V-action: the monoidal structure on
V endows V with a closed V, V-biaction.
To construct and systematically study enrichment on ∞-categories of enriched functors we extend
the concept of bitensored ∞-category to a concept of left, right and bi-enriched ∞-categories that
generalize Lurie’s notion of enriched ∞-category [16, Definition 4.2.1.28.]: Lurie defines enriched ∞-
categories as a weakening of the notion of ∞-category left tensored over a monoidal ∞-category. A
left action of a monoidal ∞-category V on an ∞-category C gives rise to a non-symmetric ∞-operad
O whose colors are the objects of V and C, and whose space of multimorphisms
MulO (V1 , ..., Vn , X, Y)
for V1 , ..., Vn ∈ V, n ≥ 0 and X, Y ∈ C is the space of maps V1 ⊗ ... ⊗ Vn ⊗ X → Y in C. For V and
C contractible we write LM for O. So LM is a non-symmetric ∞-operad with two colors a, m whose
algebras are left modules. By functoriality the ∞-operad O comes equipped with a map of ∞-operads
ON BI-ENRICHED ∞-CATEGORIES 3
O → LM, which completely determines the left action of V on C: for any objects V of V and X of C one
can reconstruct the left tensor V ⊗ X as the object corepresenting the functor MulO (V, X; −) ∶ C → S.
The left action is closed if and only if for every X, Y ∈ C there is an object MorC (X, Y) of morphisms
X → Y in C representing the presheaf MulO (−, X; Y) ∶ Vop → S. An arbitrary map of non-symmetric
∞-operads O → LM whose fibers over a and m we denote by V, C, respectively, represents a closed left
action of a closed monoidal ∞-category if the ∞-operad V is a closed monoidal ∞-category, for every
objects V ∈ V, X, Y ∈ C there is a left tensor V ⊗ X ∈ C and a morphism object MorC (X, Y) ∈ V and the
following natural maps are equivalences:
MorC (X, Y)V ≃ MorC (V ⊗ X, Y),
(1) MulO (V1 ⊗ ... ⊗ Vn , X; Y) ≃ MulO (V1 , ..., Vn , X; Y).

Lurie’s notion of enriched ∞-category arises by removing the existence of left tensors from a closed
left action: an ∞-category enriched in a monoidal ∞-category V is a LM-operad O → LM such that
for every X, Y ∈ C there is a morphism object MorC (X, Y) and equivalence (1) holds. If equivalence
(1) holds but not all morphism objects exist, Lurie calls the LM-operad O → LM a pseudo-enriched
∞-category [16, Definition 4.2.1.25.]. He calls an arbitrary LM-operad a weakly enriched ∞-category
[16, Notation 4.2.1.22.]. Pseudo-enriched and weakly enriched ∞-categories are crucial when studying
enrichment on ∞-categories of enriched functors: for a monoidal ∞-category V there is no need that the
∞-category of V-enriched presheaves is V-enriched since V itself is an example of an ∞-category of V-
enriched presheaves. But the ∞-category of presheaves is always pseudo-enriched in V. More generally,
the ∞-category of presheaves enriched in any ∞-operad V is weakly enriched in V. The concept of bi-
enriched ∞-category arises by removing the existence of bi-tensors from a closed bi-action: a bi-action
of two monoidal ∞-categories V, W on an ∞-category C gives rise to a non-symmetric ∞-operad O
whose colors are the objects of V, W, C and whose space of multimorphisms
MulO (V1 , ..., Vn , X, W1 , ..., Wm , Y)
for V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W, n, m ≥ 0 and X, Y ∈ C is the space of morphisms V1 ⊗ ... ⊗ Vn ⊗ X ⊗
W1 ⊗ ... ⊗ Wm → Y in C. The ∞-operad O comes equipped with a map of non-symmetric ∞-operads
O → BM to the ∞-operad BM governing bimodules, which completely determines the biaction: for
every objects V of V, W ∈ W and X of C the bi-tensor V ⊗ X ⊗ W corepresents the functor
MulO (V, X, W; −) ∶ C → S.
The bi-action is closed if and only if for every X, Y ∈ C the presheaf
(2) MulO (−, X, −; Y) ∶ Vop × Wop → S
is representable component-wise. We define left enriched and right enriched ∞-categories as maps
of non-symmetric ∞-operads O → BM such that for every X, Y ∈ C the presheaf (2) is representable
in the first component, in the second component, respectively, and a respective left, right pseudo-
enrichedness condition holds that is analogous to equivalence (1). We define bi-enriched ∞-categories
as maps of non-symmetric ∞-operads O → BM that are simoultaneously left and right enriched ∞-
categories. By [16, Proposition 4.8.1.17.] for every presentably monoidal ∞-categories V, W the tensor
product V ⊗ W of presentable ∞-categories identifies with the full subcategory of presheaves on V × W
that are representable in both variables. Consequently, in this situation a bi-enriched ∞-category has
morphism objects in V ⊗ W. In general, a presheaf on V × W representable in both variables uniquely
determines an adjunction V ⇄ Wop , which we denote by
LMorO (X, Y) ∶ V ⇄ Wop ∶ RMorO (X, Y).
Since pseudo-enriched and weakly enriched ∞-categories are essential when studying enrichment
on ∞-categories of enriched functors, we generalize Lurie’s notions of pseudo-enriched and weakly
enriched ∞-categories to notions of left pseudo-enriched, right pseudo-enriched and bi-pseudo-enriched
and weakly left enriched, weakly right enriched and weakly bi-enriched ∞-categories that fit into the
following diagram, which also exists for the left and right enriched cases:
{closed bi-tensored}
∞-categories
❚❚❚❚
❧❧❧❧❧ ❚❚❚❚
❧ ❧ ❚❚❚❚
❧❧ ❧❧ ❚❚❚❚
u❧❧❧ )
{bi-tensored} {∞-categories bi-enriched}
∞-categories in a monoidal ∞-category
❘❘❘
❘❘❘ ❦❦ ❦❦
❘❘❘ ❦❦❦
❘❘❘
❘❘❘ ❦❦❦❦
( u❦❦❦❦
{oplax bi-tensored} {bi-pseudo-enriched} {∞-categories bi-enriched}
∞-categories ∞-categories in an ∞-operad
❘❘❘ ❦❦
❘❘❘ ❦❦
❘❘❘ ❦❦❦
❘❘❘
❘( ❦❦❦❦❦❦❦
u❦
{weakly bi-enriched}
∞-categories
In this work we develop a theory of left, right and bi-enriched ∞-categories that studies the struc-
tures in the latter diagram and contains the theory of enriched ∞-categories in the model of Lurie as
a special case. In particular, we build a complete theory of enriched ∞-categories for Lurie’s model
of enrichment independent from the ones [3], [14].
We demonstrate that several principles of bitensored ∞-categories extend to bi-enriched ∞-categories
and their generalizations: given an ∞-category D with biaction of two monoidal ∞-categories V, W
and an ∞-category C with a left V-action the ∞-category LinFunV (C, D) of left V-linear functors
C → D, i.e. functors preserving the left V-action, carries a right W-action that applies object-wise the
right W-action on the target D. If V is a braided monoidal ∞-category, the tensor product functor
⊗ ∶ V × V → V is a monoidal functor so that every left V-action on a monoidal ∞-category D gives
rise via restriction along the tensor product of V to a left V ⊗ V-action on D. By a general principle
about the relationship between left actions and biactions a left V ⊗ V-action on D is equivalently given
by a V, Vrev -biaction on D, where Vrev is the monoidal ∞-category arising from V by reversing the
tensor product. We extend these principles about left, right and biactions to principles about left,
right and bi-enriched ∞-categories to formally treat the ∞-category of V-enriched functors like the
familiar ∞-category of V-linear functors:
Theorem 1.1. (Theorem 5.51) Let V be a locally small monoidal ∞-category and W a closed monoidal
∞-category that admits small limits. Let C be a small left V-enriched ∞-category and D a V, W-bi-
enriched ∞-category. The ∞-category
LEnrFunV (C, D)
of left V-enriched functors C → D is a right W-enriched ∞-category.
Theorem 1.2. (Theorem 4.81) Let V, W be presentably monoidal ∞-categories. There is a canonical
equivalence
V BEnrW ≃ V⊗Wrev LEnr
between the ∞-categories of V, W-bi-enriched ∞-categories and left V ⊗ Wrev -enriched ∞-categories,
where V ⊗ Wrev is the tensor product of presentable ∞-categories.
By Theorem 1.2 for any ∞-category D left enriched in a presentably braided monoidal ∞-category
V the restricted enrichment of D along the left adjoint monoidal tensor product functor ⊗ ∶ V × V → V
is an ∞-category bi-enriched in V, Vrev . So by Theorem 1.1 for any small left V-enriched ∞-category
C the ∞-category LEnrV (C, D) of left V-enriched functors C → D is a right Vrev -enriched ∞-category,
which is canonically identified with a left V-enriched ∞-category. We prove that this left V-enrichment
on the ∞-category of left V-enriched functors C → D is in fact the internal hom for the tensor product
of V-enriched ∞-categories:
Theorem 1.3. (Theorem 5.61) Let V be a presentably symmetric monoidal ∞-category and C, D small
V-enriched ∞-categories. The left V-enriched ∞-category LEnrV (C, D) provided by Theorems 1.1 and
1.2 is the internal hom for the tensor product of V-enriched ∞-categories.
Similarly, in our framework of bi-enriched ∞-categories we develop many constructions of enriched
∞-category theory via familiar constructions of the theory of left tensored, right tensored and biten-
sored ∞-categories. The ∞-category of presheaves on a small ∞-category C is the free cocompletion
under small colimits [17, Theorem 5.1.5.6.], the initial ∞-category with small colimits into which C
embeds. We construct enriched analoga of the free cocompletion. Lurie [16, Construction 2.2.4.1.]
constructs for every ∞-operad O and O-operad B an O-monoidal envelope, the initial O-monoidal ∞-
category into which B embeds. We use the free cocompletion of the BM-monoidal envelope of a small
V, W-bi-enriched ∞-category C, which we denote by PBEnv(M), to construct the free cocompletion
under small colimits and V, W-bi-tensors:
Theorem 1.4. (Theorem 4.28). Let V, W be presentably monoidal ∞-categories and C a small V, W-
bi-enriched ∞-category. There is a V, W-enriched localization PBEnv(C) → PV,W (C) such that for
every ∞-category D presentably bi-tensored over V, W the induced functor
LinFunLV,W (PV,W (C), D) → EnrFunV,W (C, D)
is an equivalence, where the left hand side is the ∞-category of left adjoint V, W-linear functors and
the right hand side is the ∞-category of V, W-enriched functors.
Specializing W to the ∞-category of spaces endowed with the cartesian structure we obtain a free
cocompletion under left tensors and small colimits, which we denote by PV (C), and dually for right
tensors by specializing V to the ∞-category of spaces: we link the universal property of free cocom-
pletion of Theorem 1.4 to another universal property of enriched presheaves: we use Theorem 1.1 to
construct for every left V-enriched ∞-category an opposite right V-enriched ∞-category (Definition
5.29, Proposition 5.45). Since any closed monoidal ∞-category V is canonically V, V-bi-enriched, by
Theorem 1.1 for any small left V-enriched ∞-category C the ∞-category REnrV (Cop , V) of right V-
enriched functors Cop → V is left V-enriched. This left V-enrichment satisfies the following universal
property: left V-enriched functors D → REnrV (Cop , V) from any left V-enriched ∞-category D corre-
spond to V, V-enriched functors D × Cop → V, where the V, V-bi-enrichment on the product D × Cop is
induced from the left V-enrichment on D and the right V-enrichment on Cop . We link this universal
property that characterizes enriched functors to the ∞-category of enriched presheaves with the uni-
versal property of free cocompletion that characterizes enriched functors starting at the ∞-category
of enriched presheaves:
Theorem 1.5. (Theorem 5.49) Let V be a presentably monoidal ∞-category and C a small left V-
enriched ∞-category. There is a left V-enriched equivalence
PV (C) ≃ REnrV (Cop , V).
By the Yoneda-lemma [17, Lemma 5.1.5.2.] every representable presheaf on a small ∞-category C
is a complete compact object [17, Definition 5.1.6.2.] of the ∞-category of presheaves P(C), an object
corepresenting a left adjoint functor P(C) → S, the functor evaluating at the representing object of
C. In enriched ∞-category theory the analogous notion of complete compact object is the notion of
atomic object [1, Definition 5.4.]: an object of a V-enriched ∞-category is atomic if it corepresents
a V-enriched left adjoint. To show Theorem 1.5 we prove that the image of any left V-enriched ∞-
category C in PV (C) and REnrV (Cop , V) consists of atomic objects. To show this for PV (C) we analyze
localizations of the free cocompletion of the enveloping left tensored ∞-category of C (§ 4). To get
control over representable V-enriched presheaves on C we prove an enriched version of the Yoneda-
lemma for left enriched and bi-enriched ∞-categories (Proposition 5.22, 5.23) via a theory of enriched
Kan extensions (Proposition 5.16, 5.20), which is of independent interest.
By Theorem 1.5 the enriched ∞-category of enriched presheaves carries two universal properties.
By the universal property of enriched cocompletion every left V-enriched functor φ ∶ C → D between
small left V-enriched ∞-categories uniquely extends to a small colimits preserving left V-linear functor
φ̄ ∶ PV (C) → PV (D). By the other universal property there is a left V-enriched functor φ∗ ∶ PV (D) →
PV (C) precomposing with φ. Theorem 1.5 implies the following corollary, which was asked by [1,
Question 1.5.]:
Corollary 1.6. (Corollary 5.50) Let V be a presentably monoidal ∞-category and φ ∶ C → D a left
V-enriched functor between small left V-enriched ∞-categories. There is a left V-enriched adjunction
φ̄ ∶ PV (C) ⇄ PV (D) ∶ φ∗ .
We use the description of the internal hom for the tensor product of enriched ∞-categories (Theorem
1.3) and the identification of enriched presheaf ∞-categories (Theorem 1.5) to compute morphism
objects in enriched ∞-categories of enriched functors. Let C be a small left V-enriched ∞-category
and D a V, W-bitensored ∞-category compatible with small colimits. Restriction to the space of
objects C≃ defines a right W-enriched functor
(3) LEnrFunV (C, D) → Fun(C≃ , D)
from left V-enriched functors C → D to functors C≃ → D, where C≃ is the maximal subspace in C. We
prove that this functor is monadic and compute the monad:
Theorem 1.7. (Theorem 5.22, Corollary 5.23) Let C be a small left V-enriched ∞-category and D
an ∞-category bitensored over V, W compatible with small colimits. The right W-enriched functor
(3) admits a right W-enriched left adjoint φ and for every functor F ∶ C≃ → D there is a canonical
equivalence of left V-enriched functors C → D ∶
colimZ∈M≃ (LMorM (Z, −) ⊗ F(Z)) ≃ φ(F).
We use Theorem 1.7 to derive monadic resolutions of left V-enriched functors and deduce the
following description of morphism objects in the right W-enriched ∞-category LEnrFunV (C, D) of left
V-enriched functors C → D ∶
Corollary 1.8. (Corollary 5.28) Let V be a locally small monoidal ∞-category and W a closed
monoidal ∞-category that admits small limits. Let C be a small left V-enriched ∞-category and D a
V, W-bi-enriched ∞-category that admits left tensors. For every left V-enriched functors F, G ∶ C → D
there is a canonical equivalence in W ∶
RMorLEnrFunV (C,D) (F, G) ≃
lim lim RMorD (LMorC (Zn−1 , Zn ) ⊗ ... ⊗ LMorC (Z1 , Z2 ) ⊗ F(Z1 ), G(Zn )).
[n]∈∆op Z1 ,....,Zn ∈C≃
In case that V is a presentably symmetric monoidal ∞-category, we prove a refinement of Corollary

5.28, which describes morphism objects in enriched functor ∞-categories via enriched ends: Gepner-
Haugseng-Nikolaus [4, Definition 2.9.] define ends for functors of ∞-categories via the twisted ar-
row ∞-category and prove that for any functors F, G ∶ C → D of ∞-categories the space of nat-
ural transformations F → F is computed as the end ∫
X∈C
mapD (F(X), G(X)). The authors define
the end of a functor H ∶ Cop × C → D as the limit of the composition Tw(C) → Cop × C Ð → D,
H
where Tw(C) → Cop × C is the twisted arrow left fibration that classifies the mapping space functor
Cop × C → S, (X, Y) ↦ mapC (X, Y). This definition characterizes the end ∫
X∈C
H as the object of D
representing the presheaf X ↦ mapFun(Cop ×C,S) (mapC , mapD (X, −) ○ H) on C. The latter universal
property generalizes to V-enriched ∞-categories for any presentably symmetric monoidal ∞-category
V by replacing mapping spaces by morphism objects: for any V-enriched ∞-categories C, D and any
V-enriched functor H ∶ Cop ⊗ C → D the V-enriched end of H is the object of D representing the
V-enriched presheaf X ↦ MorEnrFunV (Cop ⊗C,V) (MorC , MorD (X, −) ○ H) on C. We prove the following
enriched end formula:
Theorem 1.9. (Theorem 5.78) Let V be a presentably symmetric monoidal ∞-category and F, G ∶ C, D
be V-enriched functors. There is a canonical equivalence
X∈C
MorEnrFunV (C,D) (F, G) ≃ ∫ MorD ○ (Fop ⊗ G).
Closed left tensored ∞-categories and more generally Lurie’s notion of enriched ∞-categories are
typically considered to be easier to work with than Gepner-Haugseng’s model of enrichment. One
reason is that in nature enriched ∞-categories of interest often appear as full subcategories of left
tensored ∞-categories. On the other hand Gepner-Haugseng’s model has the advantage that several
elementary constructions of enriched ∞-category theory are by far easier to perform: such construc-
tions are the opposite enriched ∞-category and the transfer of enrichment: by design of the model
of Gepner-Haugseng every V-enriched ∞-category in the sense of Gepner-Haugseng can be transfered
along any lax monoidal functor φ ∶ V → V′ by applying φ to the many object associative algebra
encoding enrichment. On the other hand, for Lurie’s model of enrichment it is not clear at all how to
transfer enrichment. We construct transfer of enrichment for left, right and bi-enriched ∞-categories
via the tensored envelopes and scalar extension of presentably left, right and bi-tensored ∞-categories:
Theorem 1.10. (Theorem 4.61) Let W be a monoidal ∞-category, φ ∶ V → V′ a lax monoidal functor
and C a V, W-bi-enriched ∞-category. There is a V′ , W-bi-enriched ∞-category φ! (C) and an enriched
functor α ∶ C → φ! (C) such that
(1) the functor α is essentially surjective,
(2) for every V ∈ V, W ∈ W, X, Y ∈ C the induced morphism
φ(LMorC (X, Y)) → LMorφ! (C) (α(X), α(Y))
in V′ is an equivalence,
(3) the induced PBEnv(V′ ), PEnv(W)-linear equivalence
PBEnv(V′ ) ⊗PEnv(V) PBEnv(C) → PBEnv(φ! (C))
is an equivalence,
(4) the enriched functor α is cocartesian with respect to the forgetful functor
BEnrW → Op∞
to the ∞-category of non-symmetric ∞-operads.
We deduce the following corollary from Theorem 1.10 and Theorem 1.4:
Corollary 1.11. (Corollary 4.62) Let W be a presentably monoidal ∞-category, φ ∶ V → V′ a left
adjoint monoidal functor between presentably monoidal ∞-categories and C a V, W-bi-enriched ∞-
category. The enriched functor C → φ! (C) induces a V′ , W-linear equivalence
V′ ⊗V PV,W (C) ≃ PV,W (φ! (C)).
1.1. Notation and terminology. We fix a hierarchy of Grothendieck universes whose objects we
call small, large, very large, etc. We call a space small, large, etc. if its set of path components and
its homotopy groups are for any choice of base point. We call an ∞-category small, large, etc. if its
maximal subspace and all its mapping spaces are.
We write
● Set for the category of small sets.
● ∆ for (a skeleton of) the category of finite, non-empty, partially ordered sets and order pre-
serving maps, whose objects we denote by [n] = {0 < ... < n} for n ≥ 0.
● S for the ∞-category of small spaces.
● Cat∞ for the ∞-category of small ∞-categories.
● Catcc
∞ for the ∞-category of large ∞-categories with small colimits and small colimits preserving
functors.
We often indicate ∞-categories of large objects by (−), ̂ for example we write ̂ ̂ ∞ for the ∞-
S, Cat
categories of large spaces, ∞-categories.
For any ∞-category C containing objects A, B we write
● C(A, B) for the space of maps A → B in C,
● C/A for the ∞-category of objects over A,
● Ho(C) for its homotopy category,
● C◁ , C▷ for the ∞-category arising from C by adding an initial, final object, respectively,
● C≃ for the maximal subspace in C.
Note that Ho(Cat∞ ) is cartesian closed and for small ∞-categories C, D we write Fun(C, D) for the
internal hom, the ∞-category of functors C → D.
We often call a fully faithful functor C → D an embedding. We call a functor φ ∶ C → D an inclusion
(of a subcategory) if one of the following equivalent conditions holds:
● For any ∞-category B the induced map Cat∞ (B, C) → Cat∞ (B, D) is an embedding.
● The functor φ ∶ C → D induces an embedding on maximal subspaces and on all mapping
spaces.
Let S be an ∞-category and E ⊂ Fun([1], S) a full subcategory. We call a functor X → S a cocartesian
fibration relative to E if for every morphism [1] → S that belongs to E the pullback [1] ×S X → [1] is a
cocartesian fibration whose cocartesian morphisms are preserved by the projection [1] ×S X → X. We
call a functor X → Y over S a map of cocartesian fibrations relative to E if it preserves cocartesian lifts
of morphisms of E. We write CatE ∞/S ⊂ Cat∞/S for the subcategory of cocartesian fibrations relative to
E and maps of such. (see [11, Definition 2.3.] for details).
Acknowledgements. We thank David Gepner and Markus Spitzweck for helpful discussions. We thank
Julie Rasmusen for carefully reading a first draft of this paper.
2. Weak enrichment
In this section we recall the theory of weakly enriched ∞-categories of [10] and further develope
this theory for our needs.
2.1. Weakly enriched ∞-categories. We start with defining weakly enriched ∞-categories. To
define such we first need to introduce non-symmetric ∞-operads [3, Definition 2.2.6.], [10, Definition
2.16.].
Notation 2.1. Let Ass ∶= ∆op be the category of finite non-empty totally ordered sets and order
preserving maps. We call a map [n] → [m] in Ass
● inert if it corresponds to a map of ∆ of the form [m] ≃ {i, i + 1, ..., i + m} ⊂ [n] for some i ≥ 0.
● active if it corresponds to a map of ∆, which preserves the minimum and maximum.

Remark 2.2. For every n ≥ 0 there are n inert morphisms [n] → [1], where the i-th inert morphism
[n] → [1] for 1 ≤ i ≤ n corresponds to the map [1] ≃ {i − 1, i} ⊂ [n].
Definition 2.3. A (non-symmetric) ∞-operad is a cocartesian fibration φ ∶ V⊗ → Ass relative to the
collection of inert morphisms such that the following conditions hold, where we set V ∶= V⊗
[1]
∶
● For every n ≥ 0 the family [n] → [1] of all inert morphisms in Ass induces an equivalence
V⊗[n]
→ V×n .
● For every Y, X ∈ V⊗ lying over [m], [n] ∈ Ass the family X → Xj for 1 ≤ j ≤ n of φ-cocartesian
lifts of the inert morphisms [n] → [1] induces a pullback square
V⊗ (Y, X) / ∏1≤j≤n V⊗ (Y, Xj )

Ass([m], [n]) / ∏1≤j≤n Ass([m], [1]).
Notation 2.4. For every ∞-operad φ ∶ V⊗ → Ass we call V ∶= V⊗ [1]

the underlying ∞-category. The
∞-category V[0] is contractible. We write ∅ for the unique object.
⊗
Example 2.5. Let ∅⊗ ⊂ Ass be the full subcategory spanned by [0] ∈ Ass. Then ∅⊗ is contractible
and ∅⊗ ⊂ Ass is an ∞-operad that is the initial ∞-operad.
Notation 2.6. Let V⊗ → Ass be an ∞-operad and V1 , ..., Vn , W ∈ V for n ≥ 0. Let
MulV (V1 , ..., Vn ; W)
be the full subspace of V (V, W) spanned by the morphisms V → W in V⊗ lying over the active
⊗
morphism [1] → [n] in ∆, where V ∈ V⊗

[n]
≃ V×n corresponds to (V1 , ..., Vn ).
Definition 2.7. An ∞-operad φ ∶ V⊗ → Ass is a monoidal ∞-category if φ is a cocartesian fibration.
Remark 2.8. A cocartesian fibration φ ∶ V⊗ → Ass is a monoidal ∞-category if and only if the first
condition of Definition 2.3 holds. The second condition is automatic if φ is a cocartesian fibration [10,
Remark 2.17.].
Definition 2.9. (1) An ∞-operad φ ∶ V⊗ → Ass is locally small if the mapping spaces of V⊗ are
small.
(2) An ∞-operad φ ∶ V⊗ → Ass is small if it is locally small and V is small.
Remark 2.10. By the first axiom of Definition 2.3 an ∞-operad φ ∶ V⊗ → Ass is small if and only
if V⊗ is small. By the second axiom an ∞-operad V⊗ → Ass is locally small if and only if the multi-
morphism spaces of V⊗ → Ass are small. So an ∞-operad V⊗ → Ass is small if and only if V is small
and the multi-morphism spaces of V⊗ → Ass are small.
Notation 2.11. For every monoidal ∞-category V⊗ → Ass we write ⊗ ∶ V × V ≃ V⊗ [2]
→ V⊗
[1]
= V for
the functor induced by the unique active map [2] → [1] in Ass and we write 1V ∶ V[0] → V[1] = V for
⊗ ⊗
the functor induced by the unique active map [0] → [1] in Ass.
Definition 2.12. Let K ⊂ Cat∞ be a full subcategory.
(1) An ∞-operad V⊗ → Ass is compatible with K-indexed colimits if V admits K-indexed colimits
and for every V1 , ..., Vn , V ∈ V for n ≥ 0 and 0 ≤ i ≤ n the presheaf MulV (V1 , .., Vi , −, Vi+1 , ..., Vn ; V)
on V preserves K-indexed limits.
(2) A monoidal ∞-category is compatible with K-indexed colimits if it is compatible with K-
indexed colimits as an ∞-operad.
Remark 2.13. A monoidal ∞-category V⊗ → Ass is compatible with K-indexed colimits if and only
if V admits K-indexed colimits and such are preserved by the tensor product component-wise.
Definition 2.14. A monoidal ∞-category V⊗ → Ass is
(1) closed if the tensor product ⊗ ∶ V × V → V is left adjoint component-wise.
(2) presentably if it is closed and V is presentable.
(3) κ-compactly generated for a regular cardinal κ if V⊗ → Ass is compatible with small colimits,
V is κ-compactly generated and the monoidal structure of V restricts to the full subcategory
Vκ ⊂ V of κ-compact objects.
Remark 2.15. By [7, Proposition 7.15.] every presentably monoidal ∞-category is κ-compactly
generated for some regular cardinal κ.
We also consider maps of ∞-operads:
Definition 2.16. Let V⊗ → Ass, W⊗ → Ass be ∞-operads. A map of ∞-operads V⊗ → W⊗ is a map
of cocartesian fibrations V⊗ → W⊗ relative to the collection of inert morphisms of Ass.
Definition 2.17. A map of ∞-operads V⊗ → W⊗ is an embedding if it is fully faithful.
Definition 2.18. Let V⊗ → Ass, W⊗ → Ass be monoidal ∞-categories.
● A lax monoidal functor V⊗ → W⊗ is a map of ∞-operads V⊗ → W⊗ .
● A monoidal functor V⊗ → W⊗ is a map of cocartesian fibrations V⊗ → W⊗ over Ass.
Notation 2.19. We fix the following notation:
● Let Op∞ ⊂ Cat∞/Ass be the subcategory of ∞-operads and maps of ∞-operads.
● Let Mon ⊂ Op∞ be the subcategory of monoidal ∞-categories and monoidal functors.
● Let LMon ⊂ Mon
̂ be the subcategory of monoidal ∞-categories compatible with small colimits
and monoidal functors preserving small colimits.
Notation 2.20. For every ∞-operads V⊗ → Ass, W⊗ → Ass let
AlgV (W) ⊂ FunAss (V⊗ , W⊗ )
be the full subcategory of maps of ∞-operads. If V⊗ → Ass, W⊗ → Ass are monoidal ∞-categories, let
Fun⊗,L (V, W) ⊂ Fun⊗ (V, W) ⊂ AlgV (W)
be the full subcategories of monoidal functors (that induce on underlying ∞-categories a left adjoint).
Next we define weakly bi-enriched ∞-categories [10, Definition 3.3.].
Definition 2.21. A morphism in Ass preserves the minimum (maximum) if it corresponds to a map
[m] → [n] in ∆ sending 0 to 0 (sending m to n).
The next definition is [10, Definition 3.3.]:
Definition 2.22. Let V⊗ → Ass, W⊗ → Ass be ∞-operads. An ∞-category weakly bi-enriched in
V, W is a map φ = (φ1 , φ2 ) ∶ M⊛ → V⊗ × W⊗ of cocartesian fibrations relative to the collection of inert
morphisms of Ass × Ass whose first component preserves the maximum and whose second component
preserves the minimum such that the following conditions hold:
(1) for every n, m ≥ 0 the map [0] ≃ {n} ⊂ [n] in the first component and the map [0] ≃ {0} ⊂ [m]
in the second component induce an equivalence
θ ∶ M⊛
[n][m] → V[n] × M[0][0] × W[m] ,
⊗ ⊛ ⊗
(2) for every X, Y ∈ M⊛ lying over ([m′ ], [n′ ]), ([m], [n]) ∈ Ass × Ass the cocartesian lift Y → Y′
of the map [0] ≃ {m} ⊂ [m] and [0] ≃ {0} ⊂ [n] induces a pullback square
M⊛ (X, Y) / M⊛ (X, Y′ )

V⊗ (φ1 (X), φ1 (Y)) × W⊗ (φ2 (X), φ2 (Y)) / V⊗ (φ1 (X), φ1 (Y′ )) × W⊗ (φ2 (X), φ2 (Y′ )).
Remark 2.23. In [10, Definition 3.3.] we define weakly bi-enriched ∞-categories under the name
weakly bitensored ∞-categories and reserve the notion of weakly bi-enriched ∞-categories for a differ-
ent structure of enrichment based on Gepner-Haugseng’s model of enriched ∞-categories [3]. In [10]
we prove an equivalence between weakly bi-enriched ∞-categories and weakly bitensored ∞-categories
justifying our choice of terminology. Moreover we prefer to use this terminology since we specialize
the notion of weakly bi-enriched ∞-categories of Definition 2.22 to several notions of enrichment.
Notation 2.24. For every weakly bi-enriched ∞-category φ ∶ M⊛ → V⊗ × W⊗ we call M ∶= M⊛
[0][0]
the
underlying ∞-category of φ and say that φ exhibits M as weakly bi-enriched in V, W.
Example 2.25. Let V⊗ → Ass be an ∞-operad. We write V⊛ → Ass×Ass for the pullback of V⊗ → Ass
along the functor Ass × Ass → Ass, ([n], [m]) ↦ [n] ∗ [m]. The two functors Ass × Ass × [1] → Ass
corresponding to the natural transformations (−) ∗ ∅ → (−) ∗ (−), ∅ ∗ (−) → (−) ∗ (−) send the
morphism id[n],[m] , 0 → 1 to an inert one and so give rise to functors V⊛ → V⊗ × Ass, V⊛ → Ass × V⊗
over Ass × Ass. The resulting functor V⊛ → V⊗ × V⊗ is an ∞-category weakly bi-enriched in V, V.
Example 2.26. Let M⊛ → V⊗ ×W⊗ be a weakly bi-enriched ∞-category and N ⊂ M a full subcategory.
Let N⊛ ⊂ M⊛ be the full subcategory spanned by all objects of M⊛ lying over some (V, W) ∈ V⊗ × W⊗
corresponding some object of N ⊂ M. The restriction N⊛ ⊂ M⊛ → V⊗ × W⊗ is a weakly bi-enriched
∞-category, whose underlying ∞-category is N. We call N⊛ → V⊗ × W⊗ the full weakly bi-enriched
subcategory of M⊛ spanned by N.
Example 2.27. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category and K an ∞-category.
● The functor K × M⊛ → V⊗ × W⊗ is a weakly bi-enriched ∞-category that exhibits K × M as
weakly bi-enriched in V, W.
● The pullback (MK )⊛ ∶= V⊗ × W⊗ ×Fun(K,V⊗ ×W⊗ ) Fun(K, M⊛ ) → V⊗ × W⊗ along the diago-
nal functor V⊗ × W⊗ → Fun(K, V⊗ × W⊗ ) is a weakly bi-enriched ∞-category that exhibits
Fun(K, M) as weakly bi-enriched in V, W.
Definition 2.28. An ∞-category weakly left enriched in V is an ∞-category weakly bi-enriched in
V, ∅. An ∞-category weakly right enriched in V is an ∞-category weakly bi-enriched in ∅, V.
Notation 2.29. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category and V1 , ..., Vn ∈ V, X, Y ∈
M, W1 , ..., Wm ∈ W for some n, m ≥ 0. Let
MulM (V1 , ..., Vn , X, W1 , ..., Wm ; Y)
be the full subspace of M⊛ (Z, Y) spanned by the morphisms Z → Y in M⊛ lying over the map
[0] ≃ {0} ⊂ [n], [0] ≃ {m} ⊂ [m] in ∆ × ∆, where Z ∈ M⊛
[n],[m]
≃ V×n × M × W×m corresponds to
(V1 , ..., Vn , X, W1 , ..., Wm ).
Remark 2.30. For every ∞-operad V⊗ → Ass the projection M⊛ ∶= V⊛ → V⊗ gives rise to an
embedding V⊛ (Z, Y) → V⊗ (Z, Y) that restricts to an equivalence
MulM (V1 , ..., Vn , X, W1 , ..., Wm ; Y) ≃ MulV (V1 , ..., Vn , X, W1 , ..., Wm ; Y).
Definition 2.31. Let φ ∶ M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category.
(1) We call φ locally small if the ∞-categories V⊗ , W⊗ , M⊛ are locally small.

(2) We call φ small if φ is locally small and M is small.
(3) We call φ totally small if φ is small and V, W are small.
Remark 2.32. Let φ ∶ M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category. The first axiom of
Definition 2.22 implies that φ is totally small if and only if M⊛ , V⊗ , W⊗ are small. Remark 2.10 and
the axioms of Definition 2.22 imply that φ is locally small if and only if the multi-morphism spaces
(Notation 2.29) of V⊗ → Ass, W⊗ → Ass, φ are small. So φ is small if and only if M is small and the
multi-morphism spaces of V⊗ → Ass, W⊗ → Ass, φ are small.
2.2. Tensored ∞-categories. Next we specialize to an important class of weakly bi-enriched ∞-
categories, which is easier in nature.
The next definition is [10, Definition 3.12.]:
(1) We say that φ ∶ M⊛ → V⊗ × W⊗ exhibits M as left tensored over V if V⊗ → Ass is a monoidal
∞-category and φ is a map of cocartesian fibrations over Ass via projection to the first factor.
(2) We say that φ ∶ M⊛ → V⊗ × W⊗ exhibits M as right tensored over W if W⊗ → Ass is a
monoidal ∞-category and φ is a map of cocartesian fibrations over Ass via projection to the
second factor.
(3) We say that φ ∶ M⊛ → V⊗ × W⊗ exhibits M as bitensored over V, W if φ exhibits M as left
tensored over V and right tensored over W.
Remark 2.34. Let V⊗ → Ass, W⊗ → Ass be monoidal ∞-categories and φ ∶ M⊛ → V⊗ × W⊗ a map of
cocartesian fibrations over Ass × Ass. Then φ ∶ M⊛ → V⊗ × W⊗ exhibits M as bitensored over V, W if
and only if condition (1) of Definition 2.22 holds. Condition (2) is then automatic.
We apply Definition 2.33 in particular to the case that V⊗ = ∅⊗ or W⊗ = ∅⊗ .
Example 2.35. Let M⊛ → V⊗ be a (weakly) left tensored ∞-category and N⊛ → W⊗ a (weakly) right
tensored ∞-category. The functor M⊛ × N⊛ → V⊗ × W⊗ is a (weakly) bitensored ∞-category.
Definition 2.36. A left tensored, right tensored, bitensored ∞-category M⊛ → V⊗ ×W⊗ , respectively,
is small if M⊛ , V⊗ , W⊗ are small.
Definition 2.37. Let κ be a small regular cardinal.
(1) A left tensored ∞-category M⊛ → V⊗ × W⊗ is compatible with κ-small colimits if V⊗ →
Ass is compatible with κ-small colimits, M admits κ-small colimits, for every V ∈ V, X ∈ M
the functors (−) ⊗ X ∶ V → M, V ⊗ (−) ∶ M → M preserve κ-small colimits and for every
W1 , ..., Wm ∈ W for m ≥ 0 and Y ∈ M the functor MulM (−, W1 , ..., Wm ; Y) ∶ Mop → S preserves
κ-small limits.
(2) A right tensored ∞-category M⊛ → V⊗ × W⊗ is compatible with κ-small colimits if (Mrev )⊛ →
(Wrev )⊛ × (Vrev )⊛ is compatible with κ-small colimits.
(3) A bitensored ∞-category is compatible with κ-small colimits if it is a left and right tensored
∞-category compatible with κ-small colimits, i.e. V⊗ → Ass, W⊗ → Ass are compatible with
κ-small colimits, M admits κ-small colimits and for every V ∈ V, W ∈ W, X ∈ M the functors
(−) ⊗ X ∶ V → M, X ⊗ (−) ∶ W → M, V ⊗ (−), (−) ⊗ W ∶ M → M preserve κ-small colimits.
Definition 2.38. Let κ be a small regular cardinal.
(1) A presentably left tensored ∞-category is a left tensored ∞-category φ ∶ M⊛ → V⊗ × W⊗
compatible with small colimits such that V, M are presentable.
(2) A presentably right tensored ∞-category is a right tensored ∞-category φ ∶ M⊛ → V⊗ × W⊗
compatible with small colimits such that M, W are presentable.
(3) A presentably bitensored ∞-category is a presentably left tensored and presentably right
tensored ∞-category.
(4) A left tensored ∞-category M⊛ → V⊗ × W⊗ is κ-compactly generated if V⊗ → Ass and M are
κ-compactly generated and the left V-action on M restricts to a left Vκ -action on Mκ .
(5) A right tensored ∞-category M⊛ → V⊗ ×W⊗ is κ-compactly generated if W⊗ → Ass and M are
κ-compactly generated and the right W-action on M restricts to a right Wκ -action on Mκ .
(6) A bitensored ∞-category M⊛ → V⊗ × W⊗ is κ-compactly generated if the left V-action and
right W-action are κ-compactly generated.
By [7, Proposition 7.15.] every presentably left tensored, right tensored, bitensored ∞-category,
respectively, is κ-compactly generated for some regular cardinal κ.
Next we generalize the notions of left, right and bitensored ∞-categories (compatible with small
colimits) by introducing the notions of left and right tensors and conical colimits.
Definition 2.39. Let M⊛ → V⊗ ×W⊗ be a weakly bi-enriched ∞-category and V ∈ V, W ∈ W, X, Y ∈ M.
(1) A multi-morphism ψ ∈ MulM (V, X; Y) exhibits Y as the left tensor of V, X if for every Z ∈
M, V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0 the following map is an equivalence:
MulM (V1 , ..., Vn , Y, W1 , ..., Wm ; Z) → MulM (V1 , ..., Vn , V, X, W1 , ..., Wm ; Z).
We write V ⊗ X for Y.
(2) A multi-morphism ψ ∈ MulM (X, W; Y) exhibits Y as the right tensor of X, W if for every
Z ∈ M, V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0 the following map is an equivalence:
MulM (V1 , ..., Vn , Y, W1 , ..., Wm ; Z) → MulM (V1 , ..., Vn , X, W, W1 , ..., Wm ; Z).
We write X ⊗ W for Y.
(3) A multi-morphism τ ∈ MulM (V, Y; X) exhibits Y as the left cotensor of V, X if for every
MulM (V1 , ..., Vn , Z, W1 , ..., Wm ; Y) → MulM (V, V1 , ..., Vn , Z, W1 , ..., Wm ; X).
We write V X for Y.
(4) A multi-morphism τ ∈ MulM (Y, W; X) exhibits Y as the right cotensor of X, W if for every
MulM (V1 , ..., Vn , Z, W1 , ..., Wm ; Y) → MulM (V1 , ..., Vn , Z, W1 , ..., Wm , W; X).
We write XW for Y.
Remark 2.40. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category, V ∈ V, W ∈ W and X ∈ M. If
the respective left and right tensors exist, there is a canonical equivalence (V ⊗ X) ⊗ W ≃ V ⊗ (X ⊗ W).
In this case we refer to (V⊗X)⊗W ≃ V⊗(X⊗W) as the bitensor of V, X, W. Similarly, if the respective
left and right cotensors exist, there is a canonical equivalence (XV )W ≃ (XW )V and we refer to the
latter object as the bicotensor of V, X, W.
Definition 2.41. A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ admits
● left (co)tensors if for every object V ∈ V, X ∈ M there is a left (co)tensor of V and X.
● right (co)tensors if for every object W ∈ W, X ∈ M there is a right (co)tensor of X and W.
Example 2.42. Every weakly bi-enriched ∞-category that is left, right, bitensored admits left tensors,
right tensors, left and right tensors, respectively.
Remark 2.43. Let φ ∶ M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category that admits left and right
tensors. Then φ is a locally cocartesian fibration.
Next we define conical colimits and conical limits:
Definition 2.44. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category.

(1) Let F ∶ K▷ → M be a functor. We say that F is a conical colimit diagram if for every V1 , ..., Vn ∈
V, W1 , ..., Wm ∈ W for n, m ≥ 0 and Y ∈ M the presheaf MulM (V1 , ..., Vn , −, W1 , ..., Wm ; Y) on
M sends Fop ∶ (Kop )◁ → Mop to a limit diagram. We say that M admits K-indexed conical
colimits if M admits the conical colimit of every functor starting at K.
(2) Let F ∶ K◁ → M be a functor. We say that F is a conical limit diagram if for every V1 , ..., Vn ∈
V, W1 , ..., Wm ∈ W for n, m ≥ 0 and X ∈ M the presheaf MulM (V1 , ..., Vn , X, W1 , ..., Wm ; −) on
M sends F to a limit diagram. We say that M admits K-indexed conical limits if M admits
the conical limit of every functor starting at K.
Remark 2.45. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category.
(1) Every conical colimit diagram is a colimit diagram.
(2) If M⊛ → V⊗ × W⊗ admits left and right cotensors, every colimit diagram is conical.
(3) Assume that M⊛ → V⊗ × W⊗ admits left tensors and let G ∶ K → M be a functor. The colimit
of G is conical if and only if forming left tensors preserves the colimit of G and for every
W1 , ..., Wm ∈ W for m ≥ 0 and Y ∈ M the presheaf MulM (−, W1 , ..., Wm ; Y) on M preserves
the limit of Gop .
(4) Assume that M⊛ → V⊗ × W⊗ has left and right tensors and let G ∶ K → M be a functor. The
colimit of G is conical if and only if forming left and right tensors preserves the colimit of G.
Definition 2.46. Let K ⊂ Cat∞ be a full subcategory. A weakly bi-enriched ∞-category φ ∶ M⊛ →
V⊗ × W⊗ is compatible with K-indexed colimits if φ admits K-indexed conical colimits and for every
X, Y ∈ M, V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0 and 0 ≤ i ≤ n, 0 ≤ j ≤ m the presheaves
MulM (V1 , .., Vi , −, Vi+1 , ..., Vn , X, W1 , ..., Wm ; Y)
MulM (V1 , ..., Vn , X, W1 , .., Wj , −, Wj+1 , ..., Wm ; Y)
on V, W, respectively, preserve K-indexed limits.
2.3. Enriched functors. Next we define maps of weakly bi-enriched ∞-categories. The next defini-
tion is [10, Definition 3.18., 3.19.]:
Definition 2.47. Let α ∶ V⊗ → V′⊗ , β ∶ W⊗ → W′⊗ be maps of ∞-operads and φ ∶ M⊛ → V⊗ × W⊗ , φ′ ∶
M′⊛ → V′⊗ × W′⊗ weakly bi-enriched ∞-categories. An enriched functor φ → φ′ is a commutative
square of ∞-categories over Ass × Ass
γ
M⊛ / M′⊛
φ φ′

V⊗ × W⊗ / V′⊗ × W′⊗
α×β
such that γ preserves cocartesian lifts of inert morphisms of Ass × Ass whose first component preserves
the maximum and whose second component preserves the minimum.
● An enriched functor φ → φ′ is an embedding if α, β, γ are fully faithful.
● An enriched functor φ → φ′ is left (right) linear if it preserves left (right) tensors.
● An enriched functor φ → φ′ is linear if it is left and right linear.
● An enriched functor φ → φ′ is left V-enriched if α is the identity.
● An enriched functor φ → φ′ is right W-enriched if β is the identity.
● An enriched functor φ → φ′ is V, W-enriched if it is left V-enriched and right W-enriched.
● An enriched functor φ → φ′ is left V-linear if it is left linear and left V-enriched.
● An enriched functor φ → φ′ is right W-linear if it is right linear and right W-enriched.
● An enriched functor φ → φ′ is V, W-linear if it is left V-linear and right W-linear (or equivalently
V, W-enriched and linear).
Notation 2.48. Let M⊛ → V⊗ × W⊗ , N⊛ → V′⊗ × W′⊗ be weakly bi-enriched ∞-categories.

(1) Let
EnrFun(M, N) ⊂ (Fun(V⊗ , V′⊗ ) × Fun(W⊗ , W′⊗ )) ×Fun(M⊛ ,V′⊗ ×W′⊗ ) Fun(M⊛ , N⊛ )
be the full subcategory of enriched functors.
(2) Let
LLinFun(M, N), RLinFun(M, N), LinFun(M, N) ⊂ EnrFun(M, N)
be the full subcategories of left linear, right linear, linear functors, respectively.
Notation 2.49. Let M⊛ → V⊗ × W⊗ , N⊛ → V⊗ × W⊗ be weakly bi-enriched ∞-categories.
(1) Let
EnrFunV,W (M, N) ⊂ FunV⊗ ×W⊗ (M⊛ , N⊛ )
be the full subcategory of V, W-enriched functors.
(2) Let
LLinFunV,W (M, N) ⊂ EnrFunV,W (M, N),
RLinFunV,W (M, N) ⊂ EnrFunV,W (M, N),
LinFunV,W (M, N) ⊂ EnrFunV,W (M, N)
be the full subcategories of left linear, right linear, linear V, W-enriched functors, respectively.
(3) Let
LinFunLV,W (M, N) ⊂ LinFunV,W (M, N)
be the full subcategory of V, W-linear functors whose underlying functor admits a right adjoint.
For the next remark we use the notation of Example 2.27:
Remark 2.50. Let M⊛ → V⊗ × W⊗ , N⊛ → V⊗ × W⊗ be weakly bi-enriched ∞-categories and K an
∞-category. The canonical equivalences
Fun(K, FunV⊗ ×W⊗ (M⊛ , N⊛ )) ≃ FunV⊗ ×W⊗ (K × M⊛ , N⊛ ) ≃ FunV⊗ ×W⊗ (M⊛ , (NK )⊛ )
restrict to equivalences
Fun(K, EnrFunV,W (M, N)) ≃ EnrFunV,W (K × M, N) ≃ EnrFunV,W (M, NK ).
Let N⊛ → V′⊗ × W′⊗ be a weakly bi-enriched ∞-category. The canonical equivalence
Fun(K, (Fun(V⊗ , V′⊗ ) × Fun(W⊗ , W′⊗ )) ×Fun(M⊛ ,V′⊗ ×W′⊗ ) Fun(M⊛ , N⊛ )) ≃
(Fun(K × V⊗ , V′⊗ ) × Fun(K × W⊗ , W′⊗ )) ×Fun(K×M⊛ ,V′⊗ ×W′⊗ ) Fun(K × M⊛ , N⊛ )
restricts to an equivalence
Fun(K, EnrFun(M, N)) ≃ EnrFun(K × M, N).
We will often use the following proposition, which is [16, Proposition 4.2.4.2.]:
Proposition 2.51. For every bitensored ∞-category M⊛ → V⊗ × W⊗ and ∞-category K the forgetful
functor
LinFunV,W (V × K × W, M) → Fun(K, M)
is an equivalence.
Lemma 2.52. Let V⊗ → Ass, W⊗ → Ass be ∞-operads, M⊛ → V⊗ × W⊗ , N⊛ → V⊗ × W⊗ weakly
bi-enriched ∞-categories and K ⊂ Cat∞ a full subcategory.
(1) If N admits K-indexed conical colimits and left and right tensors, EnrFunV,W (M, N) admits
K-indexed colimits and the functor EnrFunV,W (M, N) → Fun(M, N) preserves K-indexed col-
imits.
(2) If for every V ∈ V, W ∈ W the functors V ⊗ (−), (−) ⊗ W ∶ N → N are accessible, M⊛ is small
and N is accessible, then EnrFunV,W (M, N) is accessible.
Proof. This follows from [17, Proposition 5.4.7.11., Remark 5.4.7.13.] in view of Remark 2.43.

Notation 2.53. Let

ωBEnr ⊂ (Op∞ × Op∞ ) ×Cat∞/Ass×Ass Fun([1], Cat∞/Ass×Ass )
be the subcategory of weakly bi-enriched ∞-categories.
Notation 2.54. Evaluation at the target restricts to a forgetful functor ωBEnr → Op∞ × Op∞ whose
fibers over ∞-operads V⊗ → Ass, W⊗ → Ass we denote by V ωBEnrW .
Remark 2.55. There is a canonical equivalence
ωBEnr∅,∅ ≃ Cat∞ , M⊛ → ∅⊗ × ∅⊗ ↦ M⊛
inverse to the functor K ↦ ∅⊗ × K × ∅⊗ ≃ K.
Remark 2.56. There is a canonical left action of Cat∞ on ωBEnr that sends K, M⊛ → V⊗ × W⊗ to
K × M⊛ → V⊗ × W⊗ . The forgetful functor ωBEnr → Op∞ × Op∞ is Cat∞ -linear, where the target
carries the trivial action. Thus for any small ∞-operads V⊗ → Ass, W⊗ → Ass also the fiber V ωBEnrW
carries a left Cat∞ -action that acts the same. These actions are closed by Remark 2.50 and so exhibit
ωBEnr and V ωBEnrW as left Cat∞ -enriched, i.e. as (∞, 2)-categories.
Proposition 2.57. For every small ∞-operads V⊗ → Ass, W⊗ → Ass the ∞-category V ωBEnrW is
compactly generated.
Proof. Let Catmax,min
∞/V⊗ ×W⊗
⊂ Cat∞/V⊗ ×W⊗ be the subcategory of cocartesian fibrations relative to the
collection of pairs of inert morphisms whose first component preserves the maximum and whose second
component preserves the minimum, and functors over V⊗ × W⊗ preserving cocartesian lifts of inert
morphisms whose first component preserves the maximum and whose second component preserves
the minimum. By definition V ωBEnrW ⊂ Cat∞/V⊗ ×W⊗ is a full subcategory of Catmax,min ∞/V⊗ ×W⊗
, which is
closed under filtered colimits since filtered colimits commute with finite limits. Let P be the collection
of functors [1] = [0]◁ → V⊗ selecting an inert morphism of V⊗ × W⊗ lying over a morphism of the form
[0] ≃ {n} ⊂ [n] in ∆ in the first component and a morphism of the form [0] ≃ {0} ⊂ [n] in ∆ in the
second component. By definition V ωBEnrW is the full subcategory of Catmax,min∞/V⊗ ×W⊗
spanned by the P-
fibered objects in the sense of [16, Definition B.0.19.], which are the local objects of a localization [16,
Proposition B.2.9.]. By [16, Theorem B.0.20.] the ∞-category Catmax,min ∞/V⊗ ×W⊗
is compactly generated
and so also the accessible localization V ωBEnrW is compactly generated.

Notation 2.58. Let

LMod, RMod, BMod ⊂ ωBEnr
be the subcategories of left tensored, right tensored, bitensored ∞-categories and enriched functors
preserving left tensors, right tensors, left and right tensors, respectively.
The next lemma is [10, Proposition 3.32.] whose short proof we include for the reader’s convenience:
Lemma 2.59. The forgetful functor
γ ∶ ωBEnr → Op∞ × Op∞
is a cartesian fibration that restricts to a cartesian fibration BMod → Mon × Mon with the same
cartesian morphisms. An enriched functor corresponding to a commutative square
ψ
M⊛ / N⊛

V⊗ × V′⊗ / W⊗ × W′⊗
α×α′
is γ-cartesian if and only if the square is a pullback square, i.e. if ψ induces an equivalence M ≃ N
and for any V1 , ..., Vn ∈ V, X, Y ∈ M, W1 , ..., Wm ∈ W for n, m ≥ 0 the following map is an equivalence:
MulM (V1 , ..., Vn , X, W1 , ..., Wm ; Y) → MulN (α(V1 ), ..., α(Vn ), ψ(X), α′ (W1 ), ..., α′ (Wm ); ψ(Y)).
Proof. Since Cat∞/Ass×Ass admits pullbacks, the functor Fun([1], Cat∞/Ass×Ass ) → Cat∞/Ass×Ass evalu-
ating at the target is a cartesian fibration whose cartesian morphisms are precisely the pullback squares
in Cat∞/Ass×Ass . Thus the pullback (Op∞ × Op∞ ) ×Cat∞/Ass×Ass Fun([1], Cat∞/Ass×Ass ) → Op∞ × Op∞ is
a cartesian fibration, which restricts to cartesian fibrations ωBEnr → Op∞ ×Op∞ , BMod → Mon×Mon
with the same cartesian morphisms.

Notation 2.60. The category Ass = ∆op carries a canonical involution sending [n] to [n] and a map
f ∶ [n] → [m] to the map [n] → [m], i ↦ m − f(n − i). The involution on Ass induces an involution on
Cat∞/Ass that restricts to an involution (−)rev on Op∞ ⊂ Cat∞/Ass . Moreover the involution on Ass
induces an involution on Ass × Ass by applying the involution on Ass to each factor and switching the
factors, which induces an involution on (Cat∞/Ass × Cat∞/Ass ) ×Cat∞/Ass×Ass Fun([1], Cat∞/Ass×Ass ) that
restricts to involutions (−)rev on ωBEnr and BMod, under which LMod corresponds to RMod.
Notation 2.61. Let ccBMod ⊂ BMod ̂ be the subcategory of bitensored ∞-categories compatible with
small colimits and linear functors preserving small colimits.
The next proposition follows from [10, Proposition 8.29., Proposition 3.42]:
Proposition 2.62. Let ψ ∶ M⊛ → N⊛ be an enriched functor from a totally small to a locally small
weakly bi-enriched ∞-category lying over maps of ∞-operads α ∶ V⊗ → V′⊗ , β ∶ W⊗ → W′⊗ . Let
O⊛ → V′′⊗ × W′′⊗ be a bitensored ∞-category compatible with small colimits and α′ ∶ V′⊗ → V′′⊗ , β ′ ∶
W′⊗ → W′′⊗ maps of ∞-operads. If α! ∶ AlgV (V′′ ) → AlgV′ (V′′ ) sends α′ ○α to α′ and β! ∶ AlgW (W′′ ) →
AlgW′ (W′′ ) sends β ′ ○ β to β ′ , the induced functor
EnrFunV′ ,W′ (N, (α′ , β ′ )∗ (O)) → EnrFunV,W (M, (α, β)∗ ((α′ , β ′ )∗ (O)))
admits a left adjoint ψ! , which is fully faithful if ψ is an embedding. For every V, W-enriched functor
F ∶ M⊛ → (α, β)∗ ((α′ , β ′ )∗ (O)) and X ∈ N there is a caonical equivalence
n m
ψ! (F)(X) ≃ colimV1 ,...,Vn,ψ(Y),W1 ,...,Wm→X ⊗ α(Vi ) ⊗ F(Y) ⊗ ⊗ β(Wj ).
i=1 j=1
2.4. Enriched adjunctions. Next we define enriched adjunctions.

Definition 2.63. An enriched adjunction is an adjunction in the (∞, 2)-category ωBEnr of Remark
2.56.
Definition 2.64. Let V⊗ → Ass, W⊗ → Ass be small ∞-operads. An V, W-enriched adjunction is an
adjunction in the (∞, 2)-category V ωBEnrW of Remark 2.56.
Remark 2.65. Let F ∶ M⊛ → N⊛ be an enriched functor lying over maps of ∞-operads α ∶ V⊗ → V′⊗ , β ∶
W⊗ → W′⊗ and G ∶ N⊛ → M⊛ an enriched functor lying over maps of ∞-operads γ ∶ V′⊗ → V⊗ , δ ∶ W′⊗ →
W⊗ and η ∶ id → G ○ F a morphism in EnrFun(M, M) lying over maps η ′ ∶ id → γ ○ α, η ′′ ∶ id → δ ○ β.
The morphism η ∶ id → G ○ F exhibits F as an enriched left adjoint of G (or G as an enriched right

adjoint of F) if η exhibits F as a left adjoint of G and η ′ exhibits α as a left adjoint of γ relative to Ass
and η ′′ exhibits β as a left adjoint of δ relative to Ass. If α, β, η ′ , η ′′ are the identities, a morphism
η ∶ id → G ○ F exhibits F as a V, W-enriched left adjoint of G if and only if it exhibits F as a left adjoint
of G relative to V⊗ × W⊗ . The analogous holds for the counit.
Notation 2.66. Let M⊛ → V⊗ × W⊗ , N⊛ → V⊗ × W⊗ be weakly bi-enriched ∞-categories. Let
EnrFunLV,W (M, N), EnrFunR
V,W (M, N) ⊂ EnrFunV,W (M, N)
be the full subcategories of V, W-enriched functors that admit a V, W-enriched left (right) adjoint.
Remark 2.67. The notion of enriched adjunction of Definition 2.64 was also discussed in [9, § 6.2.6.].
Remark 2.65 implies the following:
Remark 2.68. Let M⊛ → V⊗ × W⊗ , N⊛ → V′⊗ × W′⊗ be weakly bi-enriched ∞-categories.
(1) An enriched functor G ∶ N⊛ → M⊛ lying over maps of ∞-operads γ ∶ V′⊗ → V⊗ , δ ∶ W′⊗ → W⊗
admits an enriched left adjoint if γ, δ admit left adjoints α, β relative to Ass, respectively,
and for every X ∈ M there is an Y ∈ N and a morphism X → G(Y) in M such that for any
V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0 and Z ∈ N the following map is an equivalence:
MulN (α(V1 ), ..., α(Vn ), Y, β(W1 ), ..., β(Wm ); Z) →
MulM (γ(α(V1 )), ..., γ(α(Vn )), G(Y), δ(β(W1 )), ..., δ(β(Wm )); G(Z)) →
MulM (V1 , ..., Vn , X, W1 , ..., Wm ; G(Z)).
(2) An enriched functor F ∶ M⊛ → N⊛ lying over maps of ∞-operads α ∶ V⊗ → V′⊗ , β ∶ W⊗ → W′⊗
admits an enriched right adjoint if α, β admit right adjoints γ, δ relative to Ass, respectively,
and for every Y ∈ N there is an X ∈ M and a morphism F(X) → Y in N such that for any
V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0 and Z ∈ M the following map is an equivalence:
MulM (V1 , ..., Vn , Z, W1 , ..., Wm ; X) → MulN (α(V1 ), ..., α(Vn ), F(Z), β(W1 ), ..., β(Wm ); F(X))
→ MulN (α(V1 ), ..., α(Vn ), F(Z), β(W1 ), ..., β(Wm ); Y).
Lemma 2.69. Let G ∶ N⊛ → M⊛ be an enriched functor between weakly bi-enriched ∞-categories
that admit left and right tensors that lies over maps of ∞-operads γ ∶ V′⊗ → V⊗ , δ ∶ W′⊗ → W⊗ that
admit left adjoints α, β relative to Ass, respectively. Then G ∶ N⊛ → M⊛ admits an enriched left
adjoint if and only if the underlying functor N → M admits a left adjoint F ∶ M → N and for every
V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0 and X ∈ M the following morphism is an equivalence:
F(V1 ⊗ ... ⊗ Vn ⊗ X ⊗ W1 ⊗ ... ⊗ Wm ) → α(V1 ) ⊗ ... ⊗ α(Vn ) ⊗ F(X) ⊗ β(W1 ) ⊗ ... ⊗ β(Wm ).
Proof. For every V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0 and Z ∈ N the map
MulN (α(V1 ), ..., α(Vn ), F(X), β(W1 ), ..., β(Wm ); Z) →
MulM (γ(α(V1 )), ..., γ(α(Vn )), G(F(X)), δ(β(W1 )), ..., δ(β(Wm )); G(Z)) →
MulM (V1 , ..., Vn , X, W1 , ..., Wm ; G(Z)).
identifies with the composition
N(α(V1 ) ⊗ ... ⊗ α(Vn ) ⊗ F(X) ⊗ β(W1 ) ⊗ ... ⊗ β(Wm ), Z) → N(F(V1 ⊗ ... ⊗ Vn ⊗ X ⊗ W1 ⊗ ... ⊗ Wm ), Z)
≃ M(V1 ⊗ ... ⊗ Vn ⊗ X ⊗ W1 ⊗ ... ⊗ Wm , G(Z)).

Lemma 2.70. Let F ∶ M⊛ → N⊛ be an enriched functor that starts at a weakly bi-enriched ∞-category
that admits left and right tensors and lies over maps of ∞-operads α ∶ V⊗ → V′⊗ , β ∶ W⊗ → W′⊗ that
admit right adjoints γ, δ relative to Ass, respectively. Then F ∶ M⊛ → N⊛ admits an enriched right
adjoint if and only if the underlying functor M → N admits a right adjoint G ∶ N → M and F is linear.
Proof. For every V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0 and Z ∈ N the map
MulM (V1 , ..., Vn , X, W1 , ..., Wm ; G(Z)) → MulN (α(V1 ), ..., α(Vn ), F(X), β(W1 ), ..., β(Wm ); F(G(Z))) →
MulN (α(V1 ), ..., α(Vn ), F(X), β(W1 ), ..., β(Wm ); Z)
identifies with the composition
M(V1 ⊗ ... ⊗ Vn ⊗ X ⊗ W1 ⊗ ... ⊗ Wm , G(Z)) ≃ N(F(V1 ⊗ ... ⊗ Vn ⊗ X ⊗ W1 ⊗ ... ⊗ Wm ), Z) →
N(α(V1 ) ⊗ ... ⊗ α(Vn ) ⊗ F(X) ⊗ β(W1 ) ⊗ ... ⊗ β(Wm ), Z).

Notation 2.71. Let V⊗ → Ass, W⊗ → Ass be small ∞-operads. Let V ωBEnrLW , V ωBEnrR W ⊂ V ωBEnrW
be the subcategories with the same objects and with morphisms the V, W-enriched functors that admit
a V, W-enriched left adjoint, right adjoint, respectively.
Lemma 2.72. Let V⊗ → Ass, W⊗ → Ass be small ∞-operads. There is a canonical equivalence
L
V ωBEnrW ≃ (V ωBEnrR op
W) .
Proof. For every small ∞-category S let CatL∞/S , CatR ∞/S ⊂ Cat∞/S be the wide subcategories whose
morphisms are the functors over S that admit a left, right adjoint relative to S, respectively. There
is an equivalence CatL∞/S ≃ (CatR∞/S )
op
∶ a functor B → Cat∞/S is classified by a map X → S × B
of cocartesian fibrations over B, which is a map of bicartesian fibrations over B if and only if the
functor B → Cat∞/S lands in CatL∞/S [17, Corollary 5.2.2.5.]. A functor Bop → Cat∞/S is classified
by a map Y → S × B of cartesian fibrations over B, which is a map of bicartesian fibrations over
B if and only if the functor B → Cat∞/S lands in CatR ∞/S [17, Corollary 5.2.2.5.]. So there is an
equivalence Fun(B, Cat∞/S )≃ ≃ Fun(Bop , CatR∞/S )≃
≃ Fun(B, (CatR op ≃
∞/S ) ) representing an equivalence
CatL∞/S ≃ (CatR op L R
∞/S ) . The equivalence Cat∞/V⊗ ×W⊗ ≃ (Cat∞/V⊗ ×W⊗ )
op
L R op
V ωBEnr W ≃ (V ωBEnr W ) .

Remark 2.73. Let O⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category, F ∶ M⊛ → N⊛ , G ∶ N⊛ → M⊛

be V, W-enriched functors and η ∶ id → G ○ F a morphism in EnrFunV,W (M, M) that exhibits F as a
V, W-enriched left adjoint of G. Then
EnrFunV,W (η, O) ∶ EnrFunV,W (F, O) ○ EnrFunV,W (G, O) → id
exhibits EnrFunV,W (G, O) as a left adjoint of EnrFunV,W (F, O) and
EnrFunV,W (O, η) ∶ id → EnrFunV,W (O, G) ○ EnrFunV,W (O, F)
exhibits EnrFunV,W (O, F) as a left adjoint of EnrFunV,W (O, G).
2.5. Trivial weak enrichment.
Definition 2.74. Let V⊗ → Ass, W⊗ → Ass be ∞-operads.
(1) A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ is left trivial if for every n > 0, m ≥ 0
and V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W, X, Y ∈ M the space MulM (V1 , ..., Vn , X, W1 , ..., Wm ; Y) is
empty.
(2) A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ right trivial if for every n ≥ 0, m > 0

and V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W, X, Y ∈ M the space MulM (V1 , ..., Vn , X, W1 , ..., Wm ; Y) is
empty.
(3) A weakly bi-enriched ∞-category is trivial if it is left trivial and right trivial.
Remark 2.75. A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ is left trivial if and only if the
functor M⊛ → V⊗ factors through the subcategory triv⊛ ⊗
V ⊂ V . And dually for right triviality.
Remark 2.76. Let V′⊗ → V⊗ , W′⊗ → W⊗ be maps of ∞-operads. For every left trivial, right trivial,
trivial weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ the pullback V′⊗ ×V⊗ M⊛ ×W⊗ W′⊗ → V′⊗ × W′⊗
is a left trivial, right trivial, trivial weakly bi-enriched ∞-category, respectively.
Next we consider examples:
Notation 2.77. Let Assmin , Assmax ⊂ Ass be the subcategories with the same objects and with
morphisms the inert morphisms preserving the minimum, maximum, respectively.
Example 2.78. (1) The inclusion Assmax ⊂ Ass exhibits [0] as weakly left enriched in [0], where
the left enrichment is left trivial.
(2) The inclusion Assmin ⊂ Ass exhibits [0] as weakly right enriched in [0], where the right
enrichment is right trivial.
Notation 2.79. Let V⊗ → Ass, W⊗ → Ass be ∞-operads. Let
V triv
⊛
∶= V⊗ ×Ass Assmax → V⊗ , triv⊛
W ∶= W ×Ass Assmin → W ,
⊗ ⊗
V,W ∶= trivV × trivW → V × W .

triv⊛ ⊛ ⊛ ⊗ ⊗
Remark 2.76 gives the following example:

Example 2.80. Let V⊗ → Ass, W⊗ → Ass be ∞-operads. Then V triv⊛ → V⊗ , triv⊛ W → W , triv V,W →
⊗ ⊛
V × W are trivial weakly left, weakly right, weakly bi-enriched ∞-categories, respectively.
⊗ ⊗
Remark 2.81. The object [0] is a final object of the subcategories Assmax , Assmin ⊂ Ass since for
every n ≥ 0 there is a unique inert order preserving map [0] → [n] preserving the maximum, minimum,
W → Assmin are cocartesian fibrations whose fiber over
respectively. The functors V triv⊛ → Assmax , triv⊛
the final object is contractible. Thus the unique object of V triv⊛ , triv⊛
W lying over the final object is
a final object of V triv⊛ , triv⊛
W , respectively.
Definition 2.82. Let V⊗ → Ass, W⊗ → Ass be ∞-operads and K a small ∞-category. The trivial
weakly bi-enriched ∞-category on K is
V,W ∶= trivV × K × trivW → V × W .

K⊛ ⊛ ⊛ ⊗ ⊗
Remark 2.83. The weakly bi-enriched ∞-category K⊛

V,W → V × W is trivial and KV,W ≃ K.
⊗ ⊗
Proposition 2.84. Let N⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category.

(1) Let M⊛ → V⊗ be a weakly left enriched ∞-category. The following functor is an equivalence:
EnrFunV,W (M × trivW , N) → EnrFunV,∅ (M, N).
(2) Let M⊛ → W⊗ be a weakly right enriched ∞-category. The following functor is an equivalence:
EnrFunV,W (V triv × M, N) → EnrFun∅,W (M, N).
(3) The following functor is an equivalence:
EnrFunV,W (V triv × trivW , N) → N.
⊛
Proof. We prove (1). Statemement (2) is dual. (3) follows from (1) and (2). Let Ntriv ∶= N⊛ ×(V⊗ ×W⊗ )
(V⊗ × trivW ) → V⊗ × W⊗ . The embedding Ntriv
⊛ ⊛
⊂ N⊛ induces an equivalence
EnrFunV,W (M × trivW , Ntriv ) ≃ EnrFunV,W (M × trivW , N).
Consequently, it is enough to see that the canonical functor
EnrFunV,W (M × trivW , Ntriv ) → EnrFunV,∅ (M, N)
is an equivalence. The functor N⊛ → V⊗ × W⊗ over W⊗ is a map of cocartesian fibrations relative to

the subcategory triv⊛W ⊂ W . Thus the pullback Ntriv → V × trivW is map of cocartesian fibrations
⊗ ⊛ ⊗ ⊛
over trivW and so classifies a functor trivW → Cat∞/V⊗ . By Remark 2.81 the ∞-category triv⊛
⊛ ⊛
W admits
a final object so that the latter functor lifts to a functor κ ∶ trivW → (Cat∞/V⊗ )/N . This functor
⊛
⊛
[0]
sends an object W ∈ triv⊛ W to the functor λ ∶ NW → NW′ ≃ N[0] over V induced by the unique
⊛ ⊛ ⊛ ⊗
inert morphism W → W lying over the unique inert map [n] → [0] in Ass preserving the maximum.
′
By the axioms of a weakly bi-enriched ∞-category the functor λ is an equivalence so that κ is the
final object in the ∞-category of functors triv⊛
W → (Cat∞/V⊗ )/N⊛ . Hence there is an equivalence
[0]
⊛
Ntriv ⊛
≃ N[0] W over V × trivW whose pullback to ∅ ⊂ W is the identity by construction of λ.
× triv⊛ ⊗ ⊛ ⊛ ⊗
So it suffices to show that the canonical functor
EnrFunV,W (M × trivW , N × trivW ) → EnrFunV,∅ (M, N)
is an equivalence. This functor is the restriction of the composition of the canonical equivalence
W , N × trivW ) ≃ FunV⊗ ×triv⊛ (M × trivW , N × trivW )

θ ∶ FunV⊗ ×W⊗ (M⊛ × triv⊛ ⊛ ⊛ ⊛ ⊛ ⊛ ⊛
W
≃ FunV⊗ (M⊛ × triv⊛ ⊛ ⊛ ⊛ ⊛

W , N ) ≃ Fun(trivW , FunV⊗ (M , N ))
and the functor

ρ ∶ Fun(triv⊛
W , FunV⊗ (M , N )) → FunV⊗ (M , N )
⊛ ⊛ ⊛ ⊛
evaluating at the final object of triv⊛

W . The equivalence θ restricts to an equivalence
EnrFunV,W (M × trivW , N × trivW ) ≃ Fun′ (triv⊛

W , EnrFunV,∅ (M, N)),
where the right hand side is the full subcategory of Fun(triv⊛W , EnrFunV,∅ (M, N)) of functors inverting
every morphism in triv⊛W . The functor ρ restricts to an equivalence W , EnrFunV,∅ (M, N)) →
Fun′ (triv⊛
⊛
EnrFunV,∅ (M, N) since trivW admits a final object and so is weakly contractible.

Corollary 2.85. Let V⊗ → Ass, W⊗ → Ass be small ∞-operads.

(1) The functor V ωBEnrW → V ωBEnr∅ taking pullback along ∅⊛ ⊂ W⊛ admits a fully faithful
left adjoint whose essential image precisely consists of the right trivial weakly bi-enriched ∞-
categories.
(2) The functor V ωBEnrW → ∅ ωBEnrW taking pullback along ∅⊛ ⊂ V⊛ admits a fully faithful
left adjoint whose essential image precisely consists of the left trivial weakly bi-enriched ∞-
categories.
(3) The functor V ωBEnrW → Cat∞ , M⊛ → V⊗ × W⊗ ↦ M admits a fully faithful left adjoint whose
essential image precisely consists of the trivial weakly bi-enriched ∞-categories.
2.6. Tensored envelopes. In the following we introduce a tool that reduces questions about weakly
enriched ∞-categories to tensored ∞-categories. We follow [10, § 3.3.].
Notation 2.86. Let Min, Max ⊂ Fun([1], Ass) be the full subcategories of morphisms preserving the
minimum, maximum, respectively, and Ass = Min ∩ Max the intersection.
Remark 2.87. Every morphism in Ass uniquely factors as an inert morphism followed by a mor-
phism preserving the minimum and maximum. Similarly, every morphism in Ass uniquely fac-
tors as an inert morphism preserving the maximum (minimum) followed by a morphism preserv-
ing the minimum (maximum). By [17, Lemma 5.2.8.19.] this guarantees that the embeddings
Act, Min, Max ⊂ Fun([1], Ass) admit left adjoints, where a morphism in Fun([1], Ass) with local
target is a local equivalence if and only if its image in Ass under evaluation at the source is inert, is
inert and preserves the maximum, is inert and preserves the minimum, respectively, and its image in
Ass under evaluation at the target is an equivalence.
The next notation is [10, Notation 3.98.]:
Notation 2.88. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category. Let
● Env(V)⊗ ∶= Act ×Fun({0},Ass) V⊗ → Fun({1}, Ass).
● LEnv(M)⊛ ∶= Min ×Fun({0},Ass) M⊛ → Fun({1}, Ass)
● REnv(M)⊛ ∶= M⊛ ×Fun({0},Ass) Max → Fun({1}, Ass)
● BEnv(M)⊛ ∶= Min ×Fun({0},Ass) M⊛ ×Fun({0},Ass) Max → Fun({1}, Ass) × Fun([1], Ass).
Remark 2.89. The diagonal embedding Ass ⊂ Act induces embeddings
V⊗ ⊂ Env(V)⊗ , M⊛ ⊂ LEnv(M)⊛ , LEnv(M)⊛ ⊂ BEnv(M)⊛ , M⊛ ⊂ REnv(M)⊛ , REnv(M)⊛ ⊂ BEnv(M)⊛ .
Remark 2.90. Due to Remark 2.87 for any weakly bi-enriched ∞-category M⊛ → V⊗ the embeddings
Env(V)⊗ ⊂ Fun([1], Ass) ×Fun({0},Ass) V⊗ ,
LEnv(M)⊛ ⊂ Fun([1], Ass) ×Fun({0},Ass) M⊛ , REnv(M)⊛ ⊂ M⊛ ×Fun({0},Ass) Fun([1], Ass),
BEnv(M)⊛ ⊂ Fun([1], Ass) ×Fun({0},Ass) M⊛ ×Fun({0},Ass) Fun([1], Ass)
admit left adjoints relative to Fun({1}, Ass), relative to Fun({1}, Ass) × Fun({1}, Ass), respectively.
This implies that the following functors are cocartesian fibrations:
Env(V)⊗ → Fun({1}, Ass), LEnv(M)⊛ → Fun({1}, Ass), REnv(M)⊛ → Fun({1}, Ass),
BEnv(M)⊛ → Fun({1}, Ass) × Fun({1}, Ass).
Construction 2.91. There is a map
BEnv(M)⊛ ∶= Min ×Fun({0},Ass) M⊛ ×Fun({0},Ass) Max ⊂
Fun([1], Ass) ×Fun({0},Ass) M⊛ ×Fun({0},Ass) Fun([1], Ass) →
(Fun([1], Ass) ×Fun({0},Ass) V⊗ ) × (Fun([1], Ass) ×Fun({0},Ass) W⊗ ) →
Env(V)⊗ × Env(W)⊗ = (Act ×Fun({0},Ass) V⊗ ) × (Act ×Fun({0},Ass) W⊗ )
of cocartesian fibrations over Ass × Ass, where the last functor is induced by the localization functors
of Remark 2.90. There are maps
LEnv(M)⊛ ∶= Min ×Fun({0},Ass) M⊛ ⊂ Fun([1], Ass) ×Fun({0},Ass) M⊛ →
(Fun([1], Ass) ×Fun({0},Ass) V⊗ ) × W⊗ → Env(V)⊗ × W⊗ = (Act ×Fun({0},Ass) V⊗ ) × W⊗ ,
REnv(M)⊛ ∶= M⊛ ×Fun({0},Ass) Max ⊂ M⊛ ×Fun({0},Ass) Fun([1], Ass) →
V⊗ × (Fun([1], Ass) ×Fun({0},Ass) W⊗ ) → V⊗ × Env(W)⊗ = V⊗ × (Act ×Fun({0},Ass) W⊗ )
of cocartesian fibrations over Ass, where the last functors are induced by the localization functors of
Remark 2.90.

(1) The monoidal envelope of V⊗ → Ass is the functor
Env(V)⊗ → Act → Fun({1}, Ass)
(2) The left tensored envelope is the functor
LEnv(M)⊛ → Env(V)⊗ × W⊗ .
(3) The right tensored envelope is the functor
REnv(M)⊛ → REnv(M)⊛ → V⊗ × Env(W)⊗ .
(4) The bitensored envelope is the functor
BEnv(M)⊛ → Env(V)⊗ × Env(W)⊗ .
The next Proposition 2.93 justifies this terminology and is [10, Proposition 3.92, Proposition 3.101.]:
Proposition 2.93. Let V⊗ be an ∞-operad and M⊛ → V⊗ × W⊗ a weakly bi-enriched ∞-category.
(1) The functor Env(V)⊗ → Act → Fun({1}, Ass) is a monoidal ∞-category and the embedding
V⊗ ⊂ Env(V)⊗ is a map of ∞-operads. For any monoidal ∞-category W⊗ → Ass the functor
AlgEnv(V) (W) → AlgV (W) admits a fully faithful left adjoint that lands in the full subcategory
of monoidal functors. So the induced functor Fun⊗ (Env(V), W) → AlgV (W) is an equivalence.
(2) The functor LEnv(M)⊛ → Env(V)⊗ ×W⊗ is a left tensored ∞-category and the embedding M⊛ ⊂
LEnv(M)⊛ is an enriched functor. For any left tensored ∞-category N⊗ → Env(V)⊗ × W⊗
the functor EnrFunEnv(V),W(LEnv(M), N) → EnrFunV,W (M, N) admits a fully faithful left
adjoint that lands in the full subcategory of enriched functors preserving left tensors. So the
induced functor LLinFunEnv(V),W (LEnv(M), N) → EnrFunV,W (M, N) is an equivalence.
(3) The functor REnv(M)⊛ → V⊗ × Env(W)⊗ is a right tensored ∞-category and the embed-
ding M⊛ ⊂ REnv(V)⊛ is an enriched functor. For any right tensored ∞-category N⊗ →
V⊗ × Env(W)⊗ the functor EnrFunV,Env(W) (REnv(M), N) → EnrFunV,W (M, N)admits a fully
faithful left adjoint that lands in the full subcategory of enriched functors preserving right ten-
sors. So the functor RLinFunV,Env(W)(REnv(M), N) → EnrFunV,W (M, N) is an equivalence.
(4) The functor BEnv(M)⊛ → Env(V)⊗ × Env(W)⊗ is a bitensored ∞-category and the embedding
M⊛ ⊂ BEnv(M)⊛ is an enriched functor. For any bitensored ∞-category N⊗ → Env(V)⊗ ×
Env(W)⊗ the functor EnrFunEnv(V),Env(W) (BEnv(M), N) → EnrFunV,W (M, N) admits a fully
faithful left adjoint that lands in the full subcategory of linear functors. So the induced functor
LinFunV,W (BEnv(M), N) → EnrFunV,W (M, N) is an equivalence.
Remark 2.94. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category. By construction there is a
canonical V, Env(W)-linear equivalence
REnv(M)⊛ ≃ (LEnv(Mrev )rev )⊛ .
Remark 2.95. By construction every object of BEnv(M) is equivalent to V1 ⊗...⊗Vn ⊗X⊗W1 ⊗...⊗Wm
for n, m ≥ 0 and V1 , ..., Vn ∈ V ⊂ Env(V), X ∈ M ⊂ BEnv(M), W1 , ..., Wm ∈ W ⊂ Env(W). The left
tensored envelope of M⊛ → V⊗ × W⊗ is the full weakly bi-enriched subcategory
LEnv(M)⊛ ⊂ BEnv(M)⊛ ×Env(W)⊗ W⊗ → Env(V)⊗ × W⊗
spanned by the objects of the form V1 ⊗ ... ⊗ Vn ⊗ X for n ≥ 0 and V1 , ..., Vn ∈ V and X ∈ M.
Remark 2.96. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category. By construction there is a
canonical Env(V), Env(W)-linear equivalence:
LEnv(REnv(M))⊛ ≃ BEnv(M)⊛ ≃ REnv(LEnv(M))⊛ .
Remark 2.97. Let M⊛ → V⊗ be a weakly left enriched ∞-category and N⊛ → W⊗ a weakly right
enriched ∞-category. By construction there is a Env(V), Env(W)-linear equivalence:
BEnv(M × N)⊛ = Min ×Fun({0},Ass) M⊛ × N⊛ ×Fun({0},Ass) Max ≃ LEnv(M)⊛ × REnv(N)⊛ .
The next lemma is [10, Lemma 3.93.]:
Lemma 2.98. Let V⊗ → Ass be a monoidal ∞-category. The embedding V⊗ ⊂ Env(V)⊗ admits a left
adjoint relative to Ass.
The next lemma is [10, Lemma 3.105.]:
Lemma 2.99. Let M⊛ → V⊗ × W⊗ be a right tensored ∞-category. The embedding LEnv(M)⊛ ⊂
BEnv(M)⊛ admits a left Env(V)-enriched left adjoint covering the left adjoint of the embedding W⊗ ⊂
Env(W)⊗ . The enriched adjunction BEnv(M)⊛ ⇄ LEnv(M)⊛ restricts to an enriched adjunction
REnv(M)⊛ ⇄ M⊛ . So for any bitensored ∞-category M⊛ → V⊗ × W⊗ the embedding M⊛ ⊂ BEnv(M)⊛
has an enriched left adjoint covering the left adjoints of the embeddings V⊗ ⊂ Env(V)⊗ , W⊗ ⊂ Env(W)⊗ .
Remark 2.100. Let M⊛ → V⊗ × W⊗ be a right tensored ∞-category. Lemma 2.99 implies that
LEnv(M)⊛ → Env(V)⊗ × W⊗ is a bitensored ∞-category and M⊛ ⊂ LEnv(M)⊛ is right W-linear.
Next we consider the free cocompletion of the tensored envelope.
Notation 2.101. Let κ be a small regular cardinal and C a small ∞-category. Let Indκ (C) ⊂ P(C)
be the full subcategory generated by C under small κ-filtered colimits.
Example 2.102. For κ = ∅ we find that Ind∅ (C) = P(C).
Remark 2.103. By [16, Corollary 5.3.5.4.] for every small ∞-category C that admits κ-small colimits
the full subcategory Indκ (C) ⊂ P(C) precisely consists of the functors Cop → S preserving κ-small
limits. Thus Indκ (C) is a κ-accessible localization with respect to the set of maps {colim(y ○ H) →
y(colim(H)) ∣ H ∶ K → C, K κ-small}, where y ∶ C ⊂ P(C) is the Yoneda-embedding. Hence Indκ (C) is
a presentable ∞-category.
Proposition 2.104. Let κ be a small regular cardinal and V⊗ → Ass a small ∞-operad.
(1) There is an ∞-operad Indκ (V)⊗ → Ass compatible with small κ-filtered colimits and an em-
bedding V⊗ → Indκ (V)⊗ inducing the embedding V → Indκ (V) on underlying ∞-categories.
(2) For every ∞-operad W⊗ → Ass compatible with small colimits the induced functor
(4) AlgIndκ (V) (W) → AlgV (W)
admits a fully faithful left adjoint that lands in the full subcategory Algκ−fil
Indκ (V) (W) of maps of
∞-operads preserving small κ-filtered colimits. Thus the following functor is an equivalence:
(5) Indκ (V) (W) → AlgV (W).
Algκ−fil
(3) If V admits κ-small colimits, the functor (5) restricts to an equivalence
AlgLIndκ (V) (W) → AlgκV (W).
(4) If V⊗ → Ass is compatible with κ-small colimits, Indκ (V)⊗ → Ass is compatible with small
colimits.
(5) If V⊗ → Ass is a monoidal ∞-category, Indκ (V)⊗ → Ass is a monoidal ∞-category, the embed-
ding V⊗ → Indκ (V)⊗ is monoidal and for every monoidal ∞-category W⊗ → Ass compatible
with small colimits the functor (5) reflects monoidal functors.
Proof. By [10, Corollary 8.31.] for every small monoidal ∞-category V⊗ → Ass there is a monoidal
∞-category Indκ (V)⊗ → Ass compatible with small κ-filtered colimits and a monoidal embedding
V⊗ → Indκ (V)⊗ inducing the embedding V → Indκ (V) on underlying ∞-categories such that for
every monoidal ∞-category W⊗ → Ass compatible with small colimits condition (2) holds. We set
P(V)⊗ ∶= Ind∅ (V)⊗ and have a monoidal embedding V⊗ → P(V)⊗ lying over the Yoneda-embedding.
(1): Let V⊗ → Ass be a small ∞-operad and Indκ (V)⊗ ⊂ P(Env(V))⊗ the full suboperad spanned
by the essential image of the embedding Indκ (V) ⊂ P(V) ⊂ P(Env(V)). Then Indκ (V)⊗ → Ass is com-
patible with small κ-filtered colimits since P(Env(V))⊗ → Ass is a monoidal ∞-category compatible
with small colimits and the embedding Indκ (V) ⊂ P(Env(V)) preserves small κ-filtered colimits. The
embedding of ∞-operads V⊗ ⊂ Env(V)⊗ → P(Env(V))⊗ induces an embedding V⊗ → Indκ (V)⊗ .
(5): We prove that if V is a small monoidal ∞-category, both definitions of Indκ (V) agree: the
localization Env(V)⊗ → V⊗ relative to Ass of Lemma 2.98 induces a localization P(Env(V))⊗ → P(V)⊗
relative to Ass that exhibits P(V)⊗ as a full suboperad of P(Env(V))⊗ . The monoidal embedding
V⊗ → P(V)⊗ uniquely extends to a monoidal functor Indκ (V)⊗ → P(V)⊗ that preserves small κ-filtered
colimits and so induces on underlying ∞-categories the canonical embedding. So both definitions agree.
(4): Let V⊗ → Ass be a small ∞-operad compatible with κ-small colimits. We show that the
embedding
(6) Indκ (V)⊗ ⊂ P(V)⊗ ∶= Ind∅ (V)⊗
admits a left adjoint relative to Ass. This implies that Indκ (V)⊗ → Ass is compatible with small
colimits. In view of Remark 2.103 it is enough to see that for every κ-small ∞-category K, functor
H ∶ K → V and objects V1 , ..., Vn ∈ PEnv(V) for n ≥ 0 and Y ∈ Indκ (V) and 1 ≤ i ≤ n the induced map
θ ∶ MulPEnv(V) (V1 , ..., Vi , colim(H), Vi+1 , ..., Vn ; Y) → lim MulPEnv(V) (V1 , ..., Vi , H(−), Vi+1 , ..., Vn ; Y)
is an equivalence. Since PEnv(V)⊗ → Ass is a monoidal ∞-category compatible with small colimits
and PEnv(V) is generated under small colimits by tensor products of objects of V, we can moreover
assume that V1 , ..., Vn ∈ V. By [17, Corollary 5.3.5.4.] the object Y of Indκ (V) is the κ-filtered colimit
of a functor λ taking values in V, which is preserved by the κ-filtered colimits preserving embedding
Indκ (V) ⊂ P(V) and the small colimits preserving embedding P(V) ⊂ PEnv(V). The map θ factors as
MulPEnv(V) (V1 , ..., Vi , colim(H), Vi+1 , ..., Vn ; Y) ≃
colimMulPEnv(V) (V1 , ..., Vi , colim(H), Vi+1 , ..., Vn ; λ(−)) →
colim(lim MulPEnv(V) (V1 , ..., Vi , H(−), Vi+1 , ..., Vn ; λ(−))) →
lim(colimMulPEnv(V) (V1 , ..., Vi , H(−), Vi+1 , ..., Vn ; λ(−))) ≃
lim MulPEnv(V) (V1 , ..., Vi , H(−), Vi+1 , ..., Vn ; Y).
The third map in the composition commuting colimit with limit is an equivalence since κ-filtered
colimits commute with κ-small limits in the ∞-category of spaces. So for proving that θ is an equiv-
alence we can assume that Y ∈ V. In this case θ is an equivalence since V⊗ → Ass is a small ∞-operad
compatible with κ-small colimits.
(2): Let W⊗ → Ass be an ∞-operad compatible with small colimits. The following functors
̂
AlgIndκ (Env(V)) (P(Env(W))) → AlgIndκ (V) (P(Env(W))),
̂
̂
AlgIndκ (V) (P(Env(W))) → AlgV (P(Env(W)))
̂
admit fully faithful left adjoints λ, τ , respectively, by Proposition 2.62. So τ factors as λ ○ τ followed
by the first functor. By Proposition 2.93 and what we have proven the functor λ ○ τ lands in the
full subcategory of monoidal functors preserving small κ-filtered colimits. Thus τ takes values in
the full subcategory of maps of ∞-operads preserving small κ-filtered colimits and therefore sends
̂
AlgV (P(W)) ̂
to AlgIndκ (V) (P(W)) ̂
since the embedding P(W) ̂
⊂ P(Env(W)) preserves small colimits
and Indκ (V) is generated by V under small κ-filtered colimits. Consequently, τ restricts to a left
adjoint τ ′ of the induced functor
(7) Alg ̂
(P(W))
Indκ (V) → Alg (P(W)),
̂
V
which is fully faithful and lands in the full subcategory of maps of ∞-operads preserving small κ-
filtered colimits. Let σ be the strongly inaccessible cardinal corresponding to the small universe. The
localization Ind ̂
̂ σ (W)⊗ ⊂ P(W) ⊗
relative to Ass of (6) (applied to W⊗ and σ) induces two localizations
̂ σ (W)) ⊂ Alg
AlgIndκ (V) (Ind ̂ ̂ ̂
Indκ (V) (P(W)), AlgV (Indσ (W)) ⊂ AlgV (P(W)). The composed adjunction
̂
AlgV (P(W)) ⇄ AlgIndκ (V) (P(W))
̂ ⇄ AlgIndκ (V) (Ind
̂ σ (W))
restricts to an adjunction φ ∶ AlgV (Ind

̂ σ (W)) ⇄ Alg ̂
Indκ (V) (Indσ (W)). The left adjoint φ lands in the
full subcategory of maps of ∞-operads preserving small κ-filtered colimits. The functor (7) preserves
local equivalences for the respective object-wise localizations so that the unit of the latter adjunction
is a local equivalence between local objects and so an equivalence. Hence φ is fully faithful.
̂ σ (W) under small colimits and Indκ (V) is generated by V under small κ-
Since W is closed in Ind
filtered colimits, the adjunction φ ∶ AlgV (Ind
̂ σ (W)) ⇄ Alg ̂
Indκ (V) (Indσ (W)) restricts to an adjunction
AlgV (W) ⇄ AlgIndκ (V) (W), where the left adjoint lands in the full subcategory of maps of ∞-operads
preserving small κ-filtered colimits. (3) follows from [17, Proposition 5.5.1.9.].
Proposition 2.105. Let κ be a small regular cardinal and M⊛ → V⊗ × W⊗ a totally small weakly
bi-enriched ∞-category.
(1) There is a weakly bi-enriched ∞-category Indκ (M)⊛ → Indκ (V)⊗ × Indκ (W)⊗ compatible with
small κ-filtered colimits and a V, W-enriched embedding
(8) M⊛ → V⊗ ×Indκ (V)⊗ Indκ (M)⊛ ×Indκ (W)⊗ W⊗
inducing the embedding M → Indκ (M) on underlying ∞-categories.
(2) For every weakly bi-enriched ∞-category N⊛ → Indκ (V)⊗ × Indκ (W)⊗ compatible with small
colimits the functor
EnrFunIndκ (V),Indκ (W) (Indκ (M), N) → EnrFunV,W (M, N)
admits a fully faithful left adjoint that lands in the full subcategory
EnrFunκ−fil
Indκ (V),Indκ (W) (Indκ (M), N)
of enriched functors preserving small κ-filtered colimits. So the next functor is an equivalence:
(9) Indκ (V),Indκ (W) (Indκ (M), N) → EnrFunV,W (M, N).
EnrFunκ−fil
(3) If M admits κ-small colimits, equivalence (9) restricts to an equivalence
EnrFunLIndκ (V),Indκ (W) (Indκ (M), N) → EnrFunκV,W (M, N).
(4) If M⊛ → V⊗ × W⊗ is compatible with κ-small colimits, Indκ (M)⊛ → Indκ (V)⊗ × Indκ (W)⊗ is
compatible with small colimits.
(5) If M⊛ → V⊗ × W⊗ is a left tensored ∞-category, Indκ (M)⊛ → Indκ (V)⊗ × Indκ (W)⊗ is a left
tensored ∞-category, (8) is left linear and for every left tensored ∞-category N⊛ → Indκ (V)⊗ ×
Indκ (W)⊗ compatible with small colimits the functor (9) reflects left linear functors.
Proof. One proves (1)-(4) completely analogous to Proposition 2.104. By [10, Corollary 8.31.] for
every totally small weakly bitensored ∞-category M⊛ → V⊗ × W⊗ there is a bitensored ∞-category
Indκ (M)⊛ → Indκ (V)⊗ × Indκ (W)⊗ compatible with small κ-filtered colimits and a linear embedding
M⊛ → Indκ (M⊛ inducing the embedding M → Indκ (M) on underlying ∞-categories such that for
every bitensored ∞-category N⊛ → V⊗ × W⊗ compatible with small colimits condition (2) holds. We
set P(M)⊛ ∶= Ind∅ (M)⊛ → P(V)⊗ × P(W)⊗ and have a monoidal embedding M⊛ → P(M)⊛ lying over
the Yoneda-embedding.
Let M⊛ → V⊗ ×W⊗ be a totally small weakly bi-enriched ∞-category. Let Indκ (M)⊛ ⊂ P(Env(V))⊗
be the full suboperad spanned by the essential image of the embedding Indκ (M) ⊂ P(M) ⊂ P(BEnv(M)).
The linear embedding M⊛ ⊂ BEnv(M)⊛ → P(BEnv(M))⊛ induces an embedding M⊛ → Indκ (M)⊛ .
This proves (1). The proofs of (2)-(4) are analogous to Proposition 2.104. We prove (5): we prove that
if M⊛ → V⊗ × W⊗ is a left tensored ∞-category, Indκ (M)⊛ → Indκ (V)⊗ × Indκ (W)⊗ is a left tensored
∞-category and (8) is left linear. It is enough to assume that κ = ∅ because Indκ (M) is generated by
M under small κ-filtered colimits, which are preserved by the embedding Indκ (M) ⊂ P(M), so that
the left action on P(M) compatible with small colimits restricts to Indκ (M). So let κ = ∅.
By Proposition 2.99 the embedding REnv(M)⊛ ⊂ BEnv(M)⊛ admits an Env(W)-linear enriched
left adjoint lying over the left adjoint of the embedding V⊗ ⊂ Env(V)⊗ of Lemma 2.98. By (2) there
is a right P(Env(W))-linear enriched left adjoint φ ∶ P(BEnv(M))⊛ → P(REnv(M))⊛ lying over the
monoidal functor P(Env(V))⊗ → P(V)⊗ , whose underlying functor is left adjoint to the embedding
P(REnv(M)) ⊂ P(BEnv(M)). Thus the enriched right adjoint of φ is an embedding. The embed-
ding P(M)⊛ ⊂ PBEnv(M)⊛ lands in PREnv(M)⊛ and we get a left P(V)-enriched embedding M⊛ ⊂
PREnv(M)⊛ lying over P(W)⊗ ⊂ PEnv(W)⊗ . Since PREnv(M)⊛ → P(V)⊗ × PEnv(W)⊗ is a biten-
sored ∞-category compatible with small colimits and the embedding P(M) ⊂ PREnv(M) preserves
small colimits, it is enough to see that P(M) ⊂ PREnv(M) is closed under the left P(V)-action. As this
left action is compatible with small colimits, it is enough to see that M ⊂ PREnv(M) is closed under
the left V-action. This holds because the embeddings REnv(M)⊛ ⊂ V⊗ ×P(V)⊗ PREnv(M)⊛ and M⊛ ⊂
REnv(M)⊛ are left V-linear by Propositions 2.104 and 2.99. It follows that P(M)⊛ → P(V)⊗ × P(W)⊗
is a left tensored ∞-category compatible with small colimits and the embedding P(M)⊛ ⊂ PREnv(M)⊛
is left P(V)-linear. Since the embeddings M⊛ ⊂ REnv(M)⊛ and REnv(M)⊛ ⊂ V⊗ ×P(V)⊗ PREnv(M)⊛
are left V-linear, the embedding M⊛ ⊂ V⊗ ×P(V)⊗ P(M)⊛ is left V-linear, too.

Notation 2.106. Let M⊛ → V⊗ × W⊗ be a totally small weakly bi-enriched ∞-category. Let
P(V)⊗ ∶= Ind∅ (V)⊗ , P(M)⊛ ∶= Ind∅ (M)⊛ → P(V)⊗ × P(W)⊗ .
Notation 2.107. Let V⊗ → Ass, W⊗ → Ass be ∞-operads and M⊛ → V⊗ × W⊗ a weakly bi-enriched
∞-category.
(1) We call
PEnv(V)⊗ ∶= P(Env(V))⊗ → Ass
the closed monoidal envelope of V⊗ → Ass.
(2) We call
PLEnv(M)⊛ ∶= P(LEnv(M))⊛ → PEnv(V)⊗ × P(W)⊗
the closed left tensored envelope.
(3) We call
PREnv(M)⊛ ∶= P(REnv(M))⊛ → P(V)⊗ × PEnv(W)⊗
the closed right tensored envelope.
(4) We call
PBEnv(M)⊛ ∶= P(BEnv(M))⊛ → PEnv(V)⊗ × PEnv(W)⊗
the closed bi-tensored envelope.
Remark 2.108. By Proposition 2.105 the closed left tensored, right tensored, bitensored envelopes
are presentably left tensored, right tensored, bitensored ∞-categories.
Propositions 2.104, 2.105 and 2.93 give the following corollary:
Corollary 2.109. (1) Let V⊗ → Ass be a small ∞-operad. For every monoidal ∞-category W⊗ →
Ass compatible with small colimits the induced functor
AlgPEnv(V) (W) → AlgV (W)
admits a fully faithful left adjoint that lands in the full subcategory of monoidal functors that
preserve small colimits. In particular, the following induced functor is an equivalence:
Fun⊗,L (PEnv(V), W) → AlgV (W).
(2) Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category. For every bitensored ∞-category
N⊛ → PEnv(V)⊗ × PEnv(W)⊗ compatible with small colimits the induced functor
EnrFunPEnv(V),PEnv(W) (PBEnv(M), N) → EnrFunV,W (M, N)
admits a fully faithful left adjoint that lands in the full subcategory of PEnv(V), PEnv(W)-
linear functors preserving small colimits. In particular, the following functor is an equivalence:
LinFunLPEnv(V),PEnv(W) (PBEnv(M), N) → EnrFunV,W (M, N).
Remark 2.110. By the universal property of Corollary 2.109 (2) there is a canonical equivalence
PBEnv(Mrev )⊛ ≃ (PBEnv(M)rev )⊛ that restricts to an equivalence PLEnv(Mrev )⊛ ≃ (PREnv(M)rev )⊛ .
Remark 2.96 gives the following one:
Corollary 2.111. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category. There is an equivalence
PLEnv(REnv(M))⊛ ≃ PBEnv(M)⊛
of ∞-categories presentably bitensored over PEnv(V), PEnv(W).
Using tensored envelopes we can prove the following refinement of Corollary 2.72:
Proposition 2.112. Let M⊛ → V⊗ × W⊗ , N⊛ → V⊗ × W⊗ be weakly bi-enriched ∞-categories. There
is a canonical equivalence
EnrFunLV,W (M, N)op ≃ EnrFunR
V,W (N, M).
Proof. We can assume that M⊛ → V⊗ × W⊗ , N⊛ → V⊗ × W⊗ are totally small. Let K be a small

∞-category and ι ∶ M⊛ ⊂ PBEnv(M)⊛ the canonical embedding. The functor Kop × K ÐÐÐÐ→ S Ð
→
K(−,−) ι
EnrFunV,W (M, PBEnv(M)) corresponds to a V, W-enriched functor Kop × M⊛ → (PBEnv(M)⊛ )K ,

which by Corollary 2.109 corresponds to a left adjoint PEnv(V), PEnv(W)-linear functor
PBEnv(Kop × M)⊛ → (PBEnv(M)⊛ )K .
The latter is an equivalence as it induces on underlying ∞-categories the canonical equivalence
PBEnv(Kop × M) ≃ P(Kop × BEnv(M)) ≃ Fun(K, PBEnv(M)). By Corollary 2.72 there is a canonical
equivalence EnrFunLV,W (M, N)≃ ≃ EnrFunR
V,W (N, M) . Hence there is a canonical equivalence
≃
λ ∶ Cat∞ (Kop , EnrFunLPEnv(V),PEnv(W) (PBEnv(M), PBEnv(N))) ≃

Cat∞ (Kop , EnrFunV,W (M, PBEnv(N))) ≃ EnrFunV,W (Kop × M, PBEnv(N))≃ ≃
EnrFunLPEnv(V),PEnv(W) (PBEnv(M)K , PBEnv(N))≃ ≃
EnrFunR K ≃
PEnv(V),PEnv(W) (PBEnv(N), PBEnv(M) ) ≃
Cat∞ (K, EnrFunR

PEnv(V),PEnv(W) (PBEnv(N), PBEnv(M))).
Let
EnrFunLPEnv(V),PEnv(W) (PBEnv(M), PBEnv(N))′ ⊂ EnrFunLPEnv(V),PEnv(W) (PBEnv(M), PBEnv(N)),
′
EnrFunR R
PEnv(V),PEnv(W) (PBEnv(N), PBEnv(M)) ⊂ EnrFunPEnv(V),PEnv(W) (PBEnv(N), PBEnv(M))
be the full subcategories of left adjoint enriched functors sending M to N whose right adjoint sends N
to M, of right adjoint enriched functors sending N to M whose left adjoint sends M to N, respectively.
Corollary 2.109 implies that the embedding M⊛ ⊂ PBEnv(M)⊛ induces an equivalence
EnrFunLV,W (M, N) ≃ EnrFunLPEnv(V),PEnv(W) (PBEnv(M), PBEnv(N))′ .
The embedding N⊛ ⊂ PBEnv(N)⊛ induces an equivalence
′
EnrFunR R
V,W (N, M) ≃ EnrFunPEnv(V),PEnv(W) (PBEnv(N), PBEnv(M))
by Proposition 2.105 since a right adjoint functor between presheaf ∞-categories preserves small
colimits if the left adjoint preserves representables. The equivalence λ restricts to an equivalence
Cat∞ (Kop , EnrFunLPEnv(V),PEnv(W) (PBEnv(M), PBEnv(N))′ ) ≃
′
Cat∞ (K, EnrFunR
PEnv(V),PEnv(W) (PBEnv(N), PBEnv(M)) )
and we obtain an equivalence
Cat∞ (Kop , EnrFunLV,W (M, N)) ≃ Cat∞ (K, EnrFunR
V,W (N, M))
representing the desired equivalence EnrFunLV,W (M, N)op ≃ EnrFunR

V,W (N, M).

3. Enrichment and pseudo-enrichment

3.1. Pseudo-enrichment. Before defining enriched ∞-categories we define the more general class of
pseudo-enriched ∞-categories that contains all tensored ∞-categories.

(1) We say that φ exhibits M as left pseudo-enriched in V if V⊗ → Ass is a monoidal ∞-category
and for every X, Y ∈ M and V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0 the map
MulM (V1 ⊗ ... ⊗ Vn , X, W1 , ..., Wm ; Y) → MulM (V1 , ..., Vn , X, W1 , ..., Wm ; Y)
induced by the active morphism V1 , ..., Vn → V1 ⊗ ... ⊗ Vn in V⊗ is an equivalence.
(2) We say that φ exhibits M as right pseudo-enriched in W if W⊗ → Ass is a monoidal ∞-category
and for every X, Y ∈ M and V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0 the map
MulM (V1 , ..., Vn , X, W1 ⊗ ... ⊗ Wm ; Y) → MulM (V1 , ..., Vn , X, W1 , ..., Wm ; Y)
induced by the active morphism W1 , ..., Wm → W1 ⊗ ... ⊗ Wm in W⊗ is an equivalence.
(3) We say that φ exhibits M as bi-pseudo-enriched in V, W if V⊗ → Ass, W⊗ → Ass are monoidal
∞-categories and for any X, Y ∈ M and V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0 the following
map is an equivalence:
MulM (V1 ⊗ ... ⊗ Vn , X, W1 ⊗ ... ⊗ Wm ; Y) → MulM (V1 , ..., Vn , X, W1 , ..., Wm ; Y).
Remark 3.2. A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ exhibits M as bi-pseudo-enriched in
V, W if and only if it exhibits M as left pseudo-enriched in V and right pseudo-enriched in W.
Lemma 3.3. (1) A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ is a left tensored ∞-category
if and only if it is a left pseudo-enriched ∞-category and admits left tensors.
(2) A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ is a right tensored ∞-category if and only if
it is a right pseudo-enriched ∞-category and admits right tensors.
(3) A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ is a bitensored ∞-category if and only if it
is a bi-pseudo-enriched ∞-category and admits bitensors.
Proof. (2) is dual to (1) and (3) follows from (1) and (2). We prove (1): Every left tensored ∞-
category M⊛ → V⊗ × W⊗ is a left pseudo-enriched ∞-category. By left pseudo-enrichedness for every
V ∈ V, X ∈ M the natural morphism V, X → V ⊗ X in M⊛ exhibits V ⊗ X as the left tensor of V, X if
for every V′ ∈ V, Z ∈ M, W1 , ..., Wm ∈ W for m ≥ 0 the following map is an equivalence:
MulM (V′ , V ⊗ X, W1 , ..., Wm ; Z) → MulM (V′ ⊗ V, X, W1 , ..., Wm ; Z).
This follows since V′ ⊗ (V ⊗ X) ≃ (V′ ⊗ V) ⊗ X.
Conversely, if M⊛ → V⊗ × W⊗ admits left tensors, for every X ∈ M, V1 , ..., Vn ∈ V for n ≥ 0 the
functor MulM (V1 , ..., Vn , X; −) ∶ M → S is corepresentable by some object Y by iterately taking the
left tensor. For every V1′ , ..., Vk′ ∈ V, W1 , ..., Wm ∈ W, Z ∈ M for k, m ≥ 0 the canonical map
MulM (V1′ , ..., Vk′ , Y, W1 , ..., Wm ; Z) → MulM (V1′ , ..., Vk′ , V1 , ..., Vn , X, W1 , ..., Wm ; Z)
is an equivalence. By left pseudo-enrichedness for every V1j , ..., Vnj j for 1 ≤ j ≤ ℓ and W1 , ..., Wm ∈ W
for m ≥ 0 the object corepresenting the functor
MulM ({V1j , ..., Vnj j }ℓj=1 , V1 , ..., Vn , X; −) ∶ M → S
corepresents the functor MulM ({V1j ⊗ ... ⊗ Vnj j }ℓj=1 , Y; −) ∶ M → S. This implies that M⊛ → V⊗ × W⊗ is
a left tensored ∞-category.

Notation 3.4. Let

LPEnr, RPEnr, BPEnr ⊂ ωBEnr
be the full subcategories of left pseudo-enriched, right pseudo-enriched, bi-pseudo-enriched ∞-categories,
respectively.
Example 3.5. Let M⊛ → V⊗ be a weakly left enriched ∞-category and N⊛ → W⊗ a weakly right
enriched ∞-category.
(1) If M⊛ → V⊗ exhibits M as left pseudo-enriched in V, then M⊛ × N⊛ → V⊗ × W⊗ exhibits M × N
as left pseudo-enriched in V.
(2) If N⊛ → W⊗ exhibits N as right pseudo-enriched in W, then M⊛ × N⊛ → V⊗ × W⊗ exhibits
M × N as right pseudo-enriched in W.
(3) If M⊛ → V⊗ exhibits M as left pseudo-enriched in V and N⊛ → W⊗ exhibits N as right
pseudo-enriched in W, then M⊛ × N⊛ → V⊗ × W⊗ exhibits M × N as bi-pseudo-enriched in
V, W.
Construction 3.6. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category. Since PBEnv(M)⊛ →
PEnv(V)⊗ ×PEnv(W)⊗ is a closed bitensored ∞-category, by Proposition 2.51 evaluation at the tensor
units gives an equivalence
EnrFunLPEnv(V),PEnv(W) (PEnv(V) ⊗ PEnv(W), PBEnv(M)) ≃ PBEnv(M).
The graph of M is the V, W-enriched functor ΓM ∶ M⊛ × Mop → PEnv(V) ⊗ PEnv(W) corresponding
to the composition
Mop ⊂ PBEnv(M)op ≃ EnrFunLPEnv(V),PEnv(W) (PEnv(V) ⊗ PEnv(W), PBEnv(M))op
PEnv(V),PEnv(W) (PBEnv(M), PEnv(V) ⊗ PEnv(W))

≃ EnrFunR
→ EnrFunV,W (M, PEnv(V) ⊗ PEnv(W)),
where the middle functor is by Proposition 2.112 and the last functor is restriction.
Remark 3.7. For every X ∈ M the V, W-enriched functor ΓM (−, X) ∶ M⊛ → (PEnv(V) ⊗ PEnv(W))⊛
is the restriction of the enriched right adjoint of the left adjoint PEnv(V), PEnv(W)-linear functor
(PEnv(V) ⊗ PEnv(W))⊛ → PBEnv(M)⊛ , (V, W) ↦ V ⊗ X ⊗ W.
So for every X, Y ∈ M and V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0 there is an equivalence
ΓM (X, Y)(V1 ⊗ ... ⊗ Vn , W1 ⊗ ... ⊗ Wm ) ≃ PBEnv(M)(V1 ⊗ ... ⊗ Vn ⊗ X ⊗ W1 , ..., Wm , Y)
≃ MulM (V1 , ..., Vn , X, W1 , ..., Wm ; Y).

Remark 3.8. A totally small weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ exhibits M as
(1) left pseudo-enriched in V if and only if V⊗ → Ass is a monoidal ∞-category and for every
X, Y ∈ M the object
ΓM (X, Y) ∈ P(Env(V) × Env(W)) ≃ Fun(Env(W)op , PEnv(V))
lies in Fun(Env(W)op , P(V)).
(2) right pseudo-enriched in W if and only if W⊗ → Ass is a monoidal ∞-category and for every
X, Y ∈ M the object
ΓM (X, Y) ∈ P(Env(V) × Env(W)) ≃ Fun(Env(V)op , PEnv(W))
lies in Fun(Env(V)op , P(W)).
(3) bi-pseudo-enriched in V, W if and only if V⊗ → Ass, W⊗ → Ass are monoidal ∞-categories and
for every X, Y ∈ M the object
ΓM (X, Y) ∈ P(Env(V) × Env(W))
lies in P(V × W).
3.2. Enrichment. Now we are ready to define enriched ∞-categories:
(1) A left multi-morphism object of W1 , ..., Wm , X, Y ∈ M for m ≥ 0 is an object
LMulMorM (X, W1 , ..., Wn ; Y) ∈ V
such that there is a multi-morphism β ∈ MulM (LMulMorM (X, W1 , ..., Wn ; Y), X, W1 , ..., Wm ; Y)
that induces for every objects V1 , ..., Vn ∈ V for n ≥ 0 an equivalence
MulV (V1 , ..., Vn ; LMulMorM (X, W1 , ..., Wm ; Y)) ≃ MulM (V1 , ..., Vn , X, W1 , ..., Wm ; Y).
(2) A right multi-morphism object of V1 , ..., Vn , X, Y ∈ M for n ≥ 0 is an object
RMulMorM (V1 , ..., Vn , X; Y) ∈ W
such that there is a multi-morphism α ∈ MulM (V1 , ..., Vn , X, RMulMorM (V1 , ..., Vn , X; Y); Y)
that induces for every objects W1 , ..., Wm ∈ W for m ≥ 0 an equivalence
MulW (W1 , ..., Wm ; RMulMorM (V1 , ..., Vn , X; Y)) ≃ MulM (V1 , ..., Vn , X, W1 , ..., Wm ; Y).
Definition 3.10. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category and X, Y ∈ M.
(1) A left morphism object of X, Y in M is a factorization LMorM (X, Y) ∶ Env(W)op → V of the
functor Env(W)op → PEnv(V) corresponding to ΓM (X, Y) ∈ P(Env(V) × Env(W)).
(2) A right morphism object of X, Y in M is a factorization RMorM (X, Y) ∶ Env(V)op → W of the
functor Env(V)op → PEnv(W) corresponding to ΓM (X, Y) ∈ P(Env(V) × Env(W)).
Remark 3.11. Let M⊛ → V⊗ ×W⊗ be a weakly bi-enriched ∞-category and X, Y ∈ M, W1 , ..., Wm ∈ W

for m ≥ 0. A left multi-morphism object LMulMorM (W1 , ..., Wm , X; Y) ∈ V represents the presheaf
ΓM (X, Y)(W1 ⊗ ... ⊗ Wm , −) ∈ PEnv(V). Consequently, there is a left morphism object LMorM (X, Y) ∶
Env(W)op → V if and only if for every W1 , ..., Wm ∈ W for m ≥ 0 there is a left multi-morphism object
LMulMorM (W1 , ..., Wm , X; Y) ∈ V. In this case there is a canonical equivalence
LMorM (X, Y)(W1 ⊗ ... ⊗ Wm ) ≃ LMulMorM (X, W1 , ..., Wm ; Y).
The similar holds for right (multi-) morphism objects.
Lemma 3.12. Let M⊛ → V⊗ × W⊗ be a small bi-tensored ∞-category. The linear embedding M⊛ ⊂
P(M)⊛ preserves left and right multi-morphism objects.
Proof. First note that by presentability of the biaction, P(M) admits all left and right multi-morphism
objects. We prove the case of right multi-morphism objects. The case of left multi-morphism ob-
jects is similar. Let X, Y ∈ M, V1 , ..., Vn ∈ V for n ≥ 0 that admit a right multi-morphism object
RMulMorM (V1 , ..., Vn , X; Y) ∈ W in M. We like to see that the induced morphism
RMulMorM (V1 , ..., Vn , X; Y) → RMulMorP(M) (V1 , ..., Vn , X; Y)
in P(W) is an equivalence. Evaluating the latter morphism at W ∈ W gives the canonical equivalence
MulM (V1 , ..., Vn , X, W; Y) ≃ W(W, RMulMorM (V1 , ..., Vn , X; Y)) →
P(W)(W, RMulMorP(M) (V1 , ..., Vn , X; Y)) ≃ MulP(M) (V1 , ..., Vn , X, W; Y).

Proposition 3.13. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category. The embeddings
M⊛ ⊂ BEnv(M)⊛ , BEnv(M)⊛ ⊂ PBEnv(M)⊛
of weakly bi-enriched ∞-categories preserve left and right multi-morphism objects.
Proof. We first prove that the embedding ι ∶ M⊛ ⊂ BEnv(M)⊛ preserves left and right multi-morphism
objects. We prove the case of right multi-morphism objects. The case of left multi-morphism objects
is similar. Let V1 , ..., Vn ∈ V, X, Y ∈ M for n ≥ 0 and assume there is a right multi-morphism object
RMulMorM (V1 , ..., Vn , X; Y). Let jV ∶ V ⊂ Env(V), jW ∶ W ⊂ Env(W) be the canonical embeddings.
Using that BEnv(M)⊛ → Env(V)⊗ × Env(W)⊗ is a bitensored ∞-category and so a right pseudo-
enriched ∞-category, it is enough to see that for every W ∈ Env(W) the following induced map is an
equivalence:
Env(W)(W, iW (RMulMorM (V1 , ..., Vn , X; Y))) → MulPEnv(M) (jV (V1 ), ..., jV (Vn ), ι(X), W).
As W ≃ jW (W1 ) ⊗ ... ⊗ jW (Wm ) for W1 , ..., Wm ∈ W and m ≥ 0, the latter map identifies with the map
MulEnv(W) (jW (W1 ), ..., jW (Wm ), iW (RMulMorM (V1 , ..., Vn , X; Y)))
→ MulPEnv(M) (jV (V1 ), ..., jV (Vn ), ι(X), jW (W1 ), ..., jW (Wm ); ι(Y)).
The latter is an equivalence by definition of right multi-morphism object. By Lemma 3.12 the em-
bedding BEnv(M)⊛ ⊂ PBEnv(M)⊛ preserves left and right multi-morphism objects.

Definition 3.14. (1) A weakly bi-enriched ∞-category M⊛ → V⊗ ×W⊗ exhibits M as left enriched
in V if for every X, Y ∈ M there is a left morphism object LMorM (X, Y) ∈ V.
(2) A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ exhibits M as right enriched in W if for
every X, Y ∈ M there is a right morphism object RMorM (X, Y) ∈ W.
(3) A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ exhibits M as bi-enriched in V, W if it
exhibits M as left enriched in V and right enriched in W.
Notation 3.15. Let

LEnr, REnr, BEnr ⊂ ωBEnr
be the full subcategories of left enriched, right enriched, bi-enriched ∞-categories.
Example 3.16. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category whose underlying weakly
left enriched ∞-category N⊛ → V⊗ exhibits M as left V-enriched.
(1) If M⊛ → V⊗ × W⊗ admits right tensors, M⊛ → V⊗ × W⊗ exhibits M as left V-enriched. For
every X, Y ∈ M and W1 , ..., Wm for m ≥ 0 the left multi-morphism object is
LMulMorM (X, W1 , ..., Wm ; Y) = LMorN (X ⊗ W1 ⊗ ... ⊗ Wm , Y).
(2) If M⊛ → V⊗ × W⊗ admits right cotensors, M⊛ → V⊗ × W⊗ exhibits M as left V-enriched. For
every X, Y ∈ M and W1 , ..., Wm for m ≥ 0 the left multi-morphism object is
LMulMorM (X, W1 , ..., Wm ; Y) = LMorN (X, YW1 ⊗...⊗Wm ).
Remark 3.17. Let M⊛ → V⊗ × W⊗ be a bi-enriched ∞-category. The underlying weakly left enriched
∞-category N⊛ → V⊗ of M⊛ → V⊗ × W⊗ exhibits M as left V-enriched, where for any X, Y ∈ M there is
a canonical equivalence LMorN (X, Y) ≃ LMulMorM (X; Y). If M⊛ → V⊗ × W⊗ admits right tensors or
right cotensors, the left multi-morphism spaces of M⊛ → V⊗ ×W⊗ are given by the left multi-morphism
spaces of N⊛ → V⊗ by Example 3.16.
Lemma 3.18. A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ that exhibits M as left enriched in a
monoidal ∞-category, right enriched in a monoidal ∞-category, bi-enriched in monoidal ∞-categories,
respectively, exhibits M as left pseudo-enriched in V, right pseudo-enriched in W, bi-pseudo-enriched
in V, W, respectively.
Proof. We prove the case of left enrichment. The case of right enrichment is dual and the case of bi-
enrichment follows from the cases of left and right enrichment using Remark 3.2. For every X, Y ∈ M
and V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0 the map
MulM (V1 ⊗ ... ⊗ Vn , X, W1 , ..., Wm ; Y) → MulM (V1 , ..., Vn , X, W1 , ..., Wm ; Y)
induced by the active morphism V1 , ..., Vn → V1 ⊗ ... ⊗ Vn in V⊗ identifies with the induced map
MulV (V1 ⊗...⊗Vn , LMulMorM (X, W1 , ..., Wm ; Y)) → MulV (V1 , ..., Vn , LMulMorM (X, W1 , ..., Wm ; Y)),
which is an equivalence since the active morphism V1 , ..., Vn → V1 ⊗ ... ⊗ Vn is cocartesian over Ass.

Remark 3.19. Let C, D be small ∞-categories. A presheaf α ∈ P(C × D) is representable in both

variables if for every X ∈ C, Y ∈ D the presheaves α(X, −) ∶ Dop → S, α(−, Y) ∶ Cop → S are representable.
Let Prep (C × D) ⊂ P(C × D) be the full subcategory of presheaves representable in both variables.
The following conditions are equivalent for a presheaf α ∈ P(C × D):
(1) The presheaf α ∈ P(C × D) is representable in both variables.
(2) The corresponding functors Cop → P(D), Dop → P(C) induce functors Fop ∶ Cop → D, G ∶
Dop → C.
(3) The corresponding functor Dop → P(C) induces a functor G ∶ Dop → C that admits a left
adjoint. In this case for every X ∈ C the presheaf C(X, −) ○ G ≃ α(X, −) ∈ P(D) is represented
by F(X).
(4) The corresponding functor Cop → P(D) induces a functor Fop ∶ Cop → D that admits a left
adjoint. In this case for every Y ∈ D the presheaf D(Y, −)○Fop ≃ α(−, Y) ∈ P(C) is represented
by G(Y).
Consequently, the canonical equivalence P(C × D) ≃ Fun(Cop , P(D)) restricts to an equivalence

Prep (C × D) ≃ FunR (Cop , D).
The canonical equivalence P(C × D) ≃ P(D × C) restricts to an equivalence
FunR (Cop , D) ≃ Prep (C × D) ≃ Prep (D × C) ≃ FunR (Dop , C).
Let C, D be presentable ∞-categories. The universal functor C×D → C⊗D preserving small colimits
component-wise induces an embedding FunR ((C⊗D)op , S) → P(C×D), which restricts to equivalences
C ⊗ D ≃ FunR ((C ⊗ D)op , S) ≃ Prep (C × D) ≃ FunR (Cop , D)
by the adjoint functor theorem [17, Proposition 5.5.2.2.], where the first equivalence is induced by the
Yoneda-embedding. The universal functor C × D → C ⊗ D ≃ FunR (Cop , D) preserving small colimits
component-wise sends C, D to the image of D under the left adjoint lanC of the functor FunR (Cop , D) →
D evaluating at C.
Remark 3.20. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category and X, Y ∈ M such that there
is a left and a right morphism object LMorM (X, Y), RMorM (X, Y). Then
ΓM (X, Y) ∈ P(Env(V) × Env(W)) ≃ Fun(Env(V)op , PEnv(W))
is the image of RMorM (X, Y) ∈ Fun(Env(V)op , W) and
ΓM (X, Y) ∈ P(Env(V) × Env(W)) ≃ Fun(Env(W)op , PEnv(V))
is the image of LMorM (X, Y) ∈ Fun(Env(W)op , V). Thus RMorM (X, Y), LMorM (X, Y) lie in
FunR (Env(V)op , Env(W)) ≃ FunR (Env(W)op , Env(V))
so that there is an adjunction
LMorM (X, Y)op ∶ Env(W) ⇄ Env(V)op ∶ RMorM (X, Y)
whose left adjoint lands in Vop and right adjoint lands in W.
If M⊛ → V⊗ × W⊗ is bi-pseudo-enriched, ΓM (X, Y) ∈ P(Env(V) × Env(W)) lies in P(V × W), the left
morphism object LMorM (X, Y) ∶ Env(W)op → Env(V) factors as Env(W)op → Wop ÐÐÐÐÐÐÐÐÐÐ→
LMorM (X,Y)∣Wop
Env(V), the right morphism object RMorM (X, Y) ∶ Env(V)op → Env(W) factors as Env(V)op →
Vop ÐÐÐÐÐÐÐÐÐÐ→ Env(W) and there is an adjunction
RMorM (X,Y)∣Vop
LMorM (X, Y)op

∣W
∶ W ⇄ Vop ∶ RMorM (X, Y)∣V .
Remark 3.21. Let V⊗ → Ass be a monoidal ∞-category and B1⊛ ⊛
V ⊂ V the full weakly bi-enriched
subcategory spanned by the tensor unit. Then B1V → V ×V is a bi-pseudo-enriched ∞-category and
⊛ ⊗ ⊗
by the latter remark there is an adjunction V ⇄ Vop , which identifies with the canonical adjunction
LMorV (−, 1V ) ∶ V ⇄ Vop ∶ RMorV (−, 1V ) (see [12, Lemma 3.11.] for this adjunction).
Remark 3.22. Let V⊗ → Ass, W⊗ → Ass be presentably monoidal ∞-categories, M⊛ → V⊗ × W⊗ a
weakly bi-enriched ∞-category and X, Y ∈ M. Under the canonical equivalence V ⊗ W ≃ FunR (Vop , W)
the left morphism object LMorM (X, Y) is identified with an object of V ⊗ W, which we denote by
the same name. Let α be the universal monoidal functor V × W → V ⊗ W preserving small colimits
component-wise. Since for every V ∈ V, W ∈ W the object α(V, W) ∈ V ⊗ W corresponds to lanV (W) ∈
FunR (Vop , W), there is a canonical equivalence
V ⊗ W(α(V, W), LMorM (X, Y)) ≃ FunR (Vop , W)(lanV (W), LMorM (X, Y)) ≃
W(W, LMorM (X, Y)(V)) ≃ W(W, LMulMorM (V, X; Y)) ≃ MulM (V, X, W; Y).
Consequently, MulM (−, X, −; Y) ∈ P(V×W) is the image of LMorM (X, Y) ∈ V⊗W under the embedding
α∗ ∶ V ⊗ W ≃ P(V ⊗ W) → P(V × W).
Thus for every V ∈ V, W ∈ W there is a canonical equivalence

(10) V ⊗ W(α(V, W), LMorM (X, Y)) ≃ M(V ⊗ X ⊗ W, Y) ≃ MulM (V, X, W; Y).
Example 3.23. Let V⊗ → Ass, W⊗ → Ass be presentably monoidal ∞-categories and M⊛ → V⊗ × W⊗
a bitensored ∞-category such that for every X ∈ M the unique V, W-linear functor V ⊗ W → M
sending the tensor unit to X of Proposition 2.51 admits a right adjoint γX . Then M⊛ → V⊗ × W⊗
exhibits M as bi-enriched in V, W. The right adjoint sends every Y ∈ M to a left morphism object
LMorM (X, Y) ∈ V ⊗ W by equivalence (10).
Remark 3.24. By Example 3.23 for every weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ and every
X, Y ∈ M the object ΓM (X, Y) ∈ P(Env(V) × Env(W)) ≃ PEnv(V) ⊗ PEnv(W) is the left morphism
object LMorM̄ (X, Y) of X and Y, where M̄⊛ ⊂ PBEnv(M)⊛ → PEnv(V)⊗ × PEnv(W)⊗ is the full
weakly bi-enriched subcategory spanned by M.
Moreover we use the following terminology:
Definition 3.25. Let κ, τ be regular cardinals and φ ∶ M⊛ → V⊗ ×W⊗ a weakly bi-enriched ∞-category.
(1) We say that φ exhibits M as left κ-enriched in V if V⊗ → Ass is a monoidal ∞-category
compatible with κ-small colimits, φ is a left pseudo-enriched ∞-category and for every X, Y ∈ M
and W1 , ..., Wm ∈ W for m ≥ 0 the presheaf MulM (−, X, W1 , ..., Wm ; Y) on M preserves κ-small
limits.
(2) We say that φ exhibits M as right τ -enriched in W if W⊗ → Ass is a monoidal ∞-category
compatible with τ -small colimits, φ is a right pseudo-enriched ∞-category and for every X, Y ∈
M and V1 , ..., Vn ∈ V for n ≥ 0 the presheaf MulM (V1 , ..., Vn , X, −; Y) on M preserves τ -small
limits.
(3) We say that φ exhibits M as κ, τ -bi-enriched in V, W if it exhibits M as left κ-enriched in V
and right τ -enriched in W.
Notation 3.26. Let κ, τ be small regular cardinals. Let
τ
κ
LEnr, REnrτ , κ BEnr ⊂ ωBEnr
be the full subcategories of left κ-enriched, right τ -enriched, κ, τ -bi-enriched ∞-categories, respectively.
Example 3.27. A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ exhibits M as left ∅-enriched in V
if it exhibits M as left pseudo-enriched in V. A weakly bi-enriched ∞-category M⊛ → V⊗ ×W⊗ exhibits
M as right ∅-enriched in W if it exhibits M as right pseudo-enriched in W.
Example 3.28. Let κ be a small regular cardinal. A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗
is a left (right) tensored ∞-category compatible with κ-small colimits if and only if M⊛ → V⊗ × W⊗ is
a left (right) κ-enriched ∞-category that admits left (right) tensors and κ-small conical colimits.
Definition 3.29. Let σ be the strongly inaccessible cardinal corresponding to the small universe and
φ ∶ M⊛ → V⊗ × W⊗ a weakly bi-enriched ∞-category.
(1) We say that φ exhibits M as left quasi-enriched in V if φ exhibits M as left σ–enriched in V.
(2) We say that φ exhibits M as right quasi-enriched in W if φ exhibits M as right σ-enriched in
W.
(3) We say that φ exhibits M as bi-quasi-enriched in V, W if φ exhibits M as σ, σ-bi-enriched in
V, W.
Notation 3.30. Let
LQEnr, RQEnr, BQEnr ⊂ ω BEnr ̂
be the full subcategories of left quasi-enriched, right quasi-enriched, bi-quasi-enriched ∞-categories,
respectively, whose underlying ∞-category is small.
Remark 3.31. Let V⊗ → Ass, W⊗ → Ass be presentably monoidal ∞-categories. An ∞-category left
quasi-enriched in V is an ∞-category left enriched in V if and only if it is locally small. An ∞-category
right quasi-enriched in W is an ∞-category right enriched in W if and only if it is locally small. An
∞-category bi-quasi-enriched in V, W is an ∞-category bi-enriched in V, W if and only if it is locally
small.
4. Enriched presheaves via tensored envelopes

In this section we construct the ∞-category of enriched presheaves as a localization of the enveloping
∞-category with closed biaction (Proposition 4.20, Proposition 4.27, Theorem 4.28). For that we
introduce the following concept that treats weak enrichment, pseudo-enrichment and enrichment on
one footage.
Definition 4.1. A small localization pair is a pair (V⊗ → Ass, S), where V⊗ → Ass is a small ∞-operad
and S is a set of morphisms of PEnv(V) such that the saturated closure S̄ of S is closed under the
tensor product and for every V ∈ V and f ∈ S the map PEnv(V)(f, V) is an equivalence.
Remark 4.2. Let (V⊗ → Ass, S) be a small localization pair. Since PEnv(V)⊗ → Ass is a pre-
sentably monoidal ∞-category, S is a set and S̄ is closed under the tensor product, the embedding
S−1 PEnv(V)⊗ ⊂ PEnv(V)⊗ of the full suboperad spanned by the S-local objects admits a left ad-
joint relative to Ass. Moreover by definition the embedding of ∞-operads V⊗ ⊂ PEnv(V)⊗ induces an
embedding of ∞-operads V⊗ ⊂ S−1 PEnv(V)⊗ .
Definition 4.3. Let (V⊗ → Ass, S), (W⊗ → Ass, T) be small localization pairs.
(1) A weakly bi-enriched ∞-category M⊛ → V⊗ ×W⊗ exhibits M as left S-enriched in V if for every
f ∈ S, X, Y ∈ M, W ∈ PEnv(W) the induced map PBEnv(M)(f ⊗ X ⊗ W, Y) is an equivalence.
(2) A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ exhibits M as right T-enriched in W if for
every g ∈ T, X, Y ∈ M, V ∈ PEnv(V) the map PBEnv(M)(V ⊗ X ⊗ g, Y) is an equivalence.
(3) A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ exhibits M as S, T-bi-enriched in V, W if it
exhibits M as left S-enriched in V and right T-enriched in W.
Remark 4.4. A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ exhibits M as S, T-bi-enriched in
V, W if and only if for every f ∈ S̄, g ∈ T̄, X, Y ∈ M the map PBEnv(M)(f ⊗ X ⊗ g, Y) is an equivalence.
Notation 4.5. Let (V⊗ → Ass, S), (W⊗ → Ass, T) be small localization pairs. Let
S T S T
V LEnrW , V REnrW , V BEnrW ⊂ V ωBEnrW
be the full subcategories of left S-enriched, right T-enriched, S, T-bi-enriched ∞-categories, respec-
tively.
Example 4.6. Let U⊗ → Ass be a small ∞-operad. The pair (U⊗ → Ass, ∅) is a small localization
pair, where ∅−1 PEnv(U)⊗ = PEnv(U)⊗ . Let (V⊗ → Ass, S), (W⊗ → Ass, T) be small localization pairs.
Then
S S ∅ T ∅ T
V LEnrW = V BEnrW , V REnrW = V BEnrW .
Notation 4.7. Let κ be a small regular cardinal and V⊗ → Ass a small monoidal ∞-category com-
patible with κ-small colimits. Let EnrκV be the set of morphisms
V1′ , ..., Vℓ′ , V1 , ..., Vn , V1′′ , ..., Vk′′ → V1′ , ..., Vℓ′ , V1 ⊗ ... ⊗ Vn , V1′′ , ..., Vk′′ ,
V1′ , ..., Vℓ′ , colim(ι ○ F), V1′′ , ..., Vk′′ → V1′ , ..., Vℓ′ , ι(colim(F)), V1′′ , ..., Vk′′
for V1′ , ..., Vℓ′ , V1 , ..., Vn , V1′′ , ..., Vk′′ ∈ V and ℓ, n, k ≥ 0, where ι ∶ V ⊂ PEnv(V) is the embedding,
F ∶ K → V is a functor and K is κ-small.
Notation 4.8. Let κ, τ be small regular cardinals.
(1) If V⊗ → Ass is a small monoidal ∞-category compatible with κ-small colimits and W⊗ → Ass
is a small ∞-operad, let LEnrκ ∶= (EnrκV , ∅).
(2) If V⊗ → Ass is a small ∞-operad and W⊗ → Ass is a small monoidal ∞-category compatible
with κ-small colimits, let REnrκ ∶= (∅, EnrκW ).
(3) If V⊗ → Ass, W⊗ → Ass are small monoidal ∞-categories compatible with κ-small colimits,
τ -small colimits, respectively, let BEnrκ,τ ∶= (EnrκV , EnrτW ). If κ, τ are the the strongly inac-
cessible cardinal corresponding to the small universe, we drop κ, τ , respectively.
Example 4.9. Let κ be a small regular cardinal.
(1) For every small monoidal ∞-category V⊗ → Ass compatible with κ-small colimits the pair
(V⊗ → Ass, EnrκV ) is a small localization pair, where (EnrκV )−1 PEnv(V)⊗ = Indκ (V)⊗ , which is
a monoidal localization of PEnv(V)⊗ → Ass. A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗
exhibits M as left EnrκV -enriched in V if and only if it exhibits M as left κ-enriched in V.
(2) If V⊗ → Ass is a presentably monoidal ∞-category, a locally small weakly bi-enriched ∞-
category M⊛ → V⊗ × W⊗ exhibits M as left EnrV -enriched in V if and only if it exhibits M as
left enriched in V.
Definition 4.10. Let (V⊗ → Ass, S), (V′⊗ → Ass, S′ ) be small localization pairs. A map of localization
pairs (V⊗ → Ass, S) → (V′⊗ → Ass, S′ ) is a map of ∞-operads V⊗ → V′⊗ such that the induced left
adjoint monoidal functor PEnv(V) → PEnv(V′ ) sends morphisms of S to morphisms of S′ .
Example 4.11. Let V⊗ → Ass, V′⊗ → Ass be small monoidal ∞-categories compatible with κ-small
colimits, compatible with κ′ -small colimits, respectively, for small regular cardinals κ < κ′ . A map of
∞-operads V⊗ → V′⊗ is a map of localization pairs (V⊗ → Ass, EnrκV ) → (V′⊗ → Ass, EnrκV′ ) if and only
′
if it is a monoidal functor preserving κ-small colimits.

Lemma 4.12. Let C, D be presentable ∞-categories, S a set of morphisms of C and T a set of
morphisms of D. Let S̄, T̄ be the saturated classes generated by S, T and S̄ ⊗ T̄ the image of the set
S̄ × T̄ under the universal functor C × D → C ⊗ D. The left adjoint functor C ⊗ D → S−1 (C) ⊗ T−1 D
admits a fully faithful right adjoint whose essential image are the S̄ ⊗ T̄-local objects.
Proof. By [16, Proposition 4.8.1.17.] there is an equivalence C ⊗ D ≃ FunR (Cop , D), where the right
hand side is the full subcategory of right adjoint functors Cop → D. The functor C⊗D → S−1 (C)⊗T−1 D
is left adjoint to the embedding FunR (S−1 (C)op , T−1 D) ⊂ FunR (Cop , T−1 D) → FunR (Cop , D) whose
essential image are the small limits preserving functors Cop → D that land in T−1 D and invert S.
The functor C × D → C ⊗ D ≃ FunR (Cop , D) sends X, Y to Cop ÐÐÐÐ→ S ÐÐÐ→ D. An object
C(−,X) (−)⊗Y
F ∈ C ⊗ D ≃ FunR (Cop , D) is local with respect to S̄ ⊗ T̄ if and only if it is local with respect to
f ⊗ Y, X ⊗ g for any f ∈ S̄, g ∈ T̄, X ∈ C, Y ∈ D. There are canonical equivalences
C ⊗ D(f ⊗ Y, F) ≃ FunR (Cop , D)(C(−, f) ⊗ Y, F) ≃ D(Y, F(f)),
C ⊗ D(X ⊗ g, F) ≃ FunR (Cop , D)(C(−, X) ⊗ g, F) ≃ D(g, F(X)).
So F is local with respect to f ⊗ Y for any f ∈ S̄, Y ∈ D if and only if F inverts maps of S̄ (equivalently
of S). So F is local with respect to X ⊗ g for any X ∈ C, g ∈ T̄ if and only if F lands in T−1 D.

Corollary 4.13. Let (V⊗ → Ass, S), (W⊗ → Ass, T) be small localization pairs.
(1) A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ exhibits M as left S-enriched if for every
X, Y ∈ M, W1 , ..., Wm ∈ W for m ≥ 0 the left multi-morphism object
LMulMorM̄ (W1 , ..., Wm , X; Y)
−1
belongs to S PEnv(V).
(2) A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ exhibits M as right T-enriched if for every

X, Y ∈ M, V1 , ..., Vn ∈ V for n ≥ 0 the right multi-morphism object
RMulMorM̄ (V1 , ..., Vn , X; Y)
−1
belongs to T PEnv(W).
(3) A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ exhibits M as S, T-bi-enriched if for every
X, Y ∈ M, V1 , ..., Vn ∈ V for n ≥ 0 the morphism object MorM̄ (X, Y) belongs to S−1 PEnv(V) ⊗
T−1 PEnv(W).
Proof. Statements (1) and (2) are clear. We prove (3): for any X, Y ∈ PBEnv(M), f ∈ S̄, g ∈ T̄ the map
PBEnv(M)(f ⊗ X ⊗ g, Y) identifies with the map PEnv(V) ⊗ PEnv(W)(f ⊗ g, MorPBEnv(M) (X, Y)). So
the result follows from Lemma 4.12.

Notation 4.14. Let (V⊗ → Ass, S), (W⊗ → Ass, T) be small localization pairs (Definition 4.1) and
M⊛ → V⊗ × W⊗ a small weakly bi-enriched ∞-category.
(1) Let
S,T ⊂ S PEnv(V) ×PEnv(V)⊗ PBEnv(M) ×PEnv(W)⊗ T PEnv(W)

PBEnv(M)⊛ −1 ⊗ ⊛ −1 ⊗
be the full weakly bi-enriched subcategory spanned by the presheaves on BEnv(M) that are
local with respect to the set of morphisms of the form f ⊗X⊗W1 ...⊗Wm and V1 ⊗...⊗Vn ⊗X⊗g
for f ∈ S, g ∈ T, X ∈ M, V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0.
(2) Let
S,T ∶= V ×S−1 PEnv(V)⊗ PBEnv(M)S,T ×T−1 PEnv(W)⊗ W .
̃
PBEnv(M) ⊛ ⊗ ⊛ ⊗
If S = ∅ or T = ∅, we drop S, T from the notation, respectively.

Remark 4.15. As a consequence of Remark 2.110 there is a canonical equivalence
⊛ rev ⊛
(PBEnv(M)rev
S,T ) ≃ PBEnv(M )T,S .
Remark 4.16. Since PBEnv(M) is presentable, the embedding PBEnv(M)S,T ⊂ PBEnv(M) admits
a left adjoint. The full subcategory PBEnv(M)S,T ⊂ PBEnv(M) precisely consists of the presheaves on
BEnv(M) that are local with respect to the collection Q of morphisms of the form f⊗X⊗g for f ∈ S̄, g ∈ T̄
and X ∈ M. The biaction of PEnv(V), PEnv(W) on PBEnv(M) sends morphisms of S̄, Q, T̄ to Q. This
implies that the embedding PBEnv(M)⊛ ⊛
S,T ⊂ PBEnv(M) admits an enriched left adjoint such that the
unit lies over the units of the localizations S PEnv(V) ⊂ PEnv(V)⊗ , T−1 PEnv(W)⊗ ⊂ PEnv(W)⊗ .
−1 ⊗
Remark 4.17. A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ is a S, T-bi-enriched ∞-category if

and only if M⊛ ⊂ PBEnv(M)⊛ lands in PBEnv(M)⊛S,T .
Remark 4.18. Let M⊛ → N⊛ be an enriched functor between totally small weakly bi-enriched ∞-
categories lying over maps of small localization pairs
(V⊗ → Ass, S) → (V′⊗ → Ass, S′ ), (W⊗ → Ass, T) → (W′⊗ → Ass, T′ ).
The induced left adjoint linear functor PBEnv(M)⊛ → PBEnv(N)⊛ preserves local equivalences for
the corresponding localizations and so descends to a left adjoint linear functor
S,T → PBEnv(N)S′ ,T′ .
PBEnv(M)⊛ ⊛
Lemma 4.19. Let (V⊗ → Ass, S), (W⊗ → Ass, T) be localization pairs, M⊛ → V⊗ × W⊗ a totally small
weakly bi-enriched ∞-category and Y ∈ PBEnv(M). The following conditions are equivalent:
(1) The object Y belongs to PBEnv(M)S,T .
(2) For every X ∈ M the object MorPBEnv(M) (X, Y) belongs to S−1 PEnv(V) ⊗ T−1 PEnv(W).
(3) The object Y is local with respect to all morphisms f ⊗X⊗g for f ∈ S̄, g ∈ T̄ and X ∈ PBEnv(M).
(4) For any X ∈ PBEnv(M) the object MorPBEnv(M) (X, Y) belongs to S−1 PEnv(V)⊗T−1 PEnv(W).
Proof. Lemma 4.12 implies that (1) is equivalent to (2), and that (3) is equivalent to (4). Condition
(1) is equivalent to (3) since S̄, T̄ are closed under the tensor product and PBEnv(M) is generated by
M under small colimits and the PEnv(V), PEnv(W)-biaction.

Proposition 4.20. Let (V⊗ → Ass, S), (W⊗ → Ass, T) be small localization pairs, M⊛ → V⊗ × W⊗
a small S, T-bi-enriched ∞-category and ρ ∶ N⊛ → S−1 PEnv(V)⊗ × T−1 PEnv(W)⊗ a bitensored ∞-
category compatible with small colimits.
(1) The enriched embedding M⊛ ⊂ PBEnv(M)⊛
S,T induces a functor
EnrFunS−1 PEnv(V),T−1 PEnv(W)(PBEnv(M)S,T , N) → EnrFunV,W (M, N)
that admits a fully faithful left adjoint. The left adjoint lands in the full subcategory
LinFunLS−1 PEnv(V),T−1 PEnv(W) (PBEnv(M)S,T , N).
(2) The following induced functor is an equivalence:
LinFunLS−1 PEnv(V),T−1 PEnv(W) (PBEnv(M)S,T , N) → EnrFunV,W (M, N).
(3) There is a canonical S−1 PEnv(V), T−1 PEnv(W)-linear equivalence
S,T ≃ (S PEnv(V) ⊗PEnv(V) PBEnv(M) ⊗PEnv(W) T PEnv(W)) .

PBEnv(M)⊛ −1 −1 ⊛
Proof. We first prove (2). Let N′⊛ → PEnv(V)⊗ × PEnv(W)⊗ be the pullback of ρ along the left
adjoints of the embeddings S−1 PEnv(V)⊗ ⊂ PEnv(V)⊗ , T−1 PEnv(W)⊗ ⊂ PEnv(W)⊗ . Consider
LinFunLS−1 PEnv(V),T−1 PEnv(W) (PBEnv(M)S,T , N) Ð

→ LinFunLPEnv(V),PEnv(W)(PBEnv(M), N′ )
ρ
→ EnrFunV,W (M, N).

The first functor in the composition is an equivalence by the description of local equivalences. This
proves (3). The second functor is an equivalence by Propositions 2.93 and 2.104.
(1): The left adjoint of the functor
EnrFunS−1 PEnv(V),T−1 PEnv(W)(PBEnv(M)S,T , N) → EnrFunV,W (M, N)
factors as
EnrFunV,W (M, N) Ð
→ EnrFunPEnv(V),PEnv(W) (PBEnv(M), N′ ) Ð
→
α β
EnrFunS−1 PEnv(V),T−1 PEnv(W) (PBEnv(M)S,T , N),

where β takes restriction and α is the fully faithful left adjoint of
EnrFunPEnv(V),PEnv(W)(PBEnv(M), N′ ) → EnrFunV,W (M, N)
that lands in LinFunLPEnv(V),PEnv(W) (PBEnv(M), N′ ) (Proposition 2.105). Since ρ is essentially surjec-

tive, β takes LinFunLPEnv(V),PEnv(W) (PBEnv(M), N′ ) to LinFunLS−1 PEnv(V),T−1 PEnv(W)(PBEnv(M)S,T , N).

4.1. Enriched presheaves. In the following we specialize the theory of the latter subsection to
construct an enriched ∞-category of enriched presheaves and more generally for every small regular
cardinal κ a κ-enriched ∞-category of κ-enriched presheaves.
Notation 4.21. Let κ be a small regular cardinal and V⊗ → Ass a small monoidal ∞-category.
(1) Let
Pκ (V)⊗ ⊂ P(V)⊗
be the full suboperad spanned by the full subcategory of P(V) generated by V under κ-small
colimits. For κ the large strongly inaccessible cardinal corresponding to the small universe
(appyling the notation in a larger universe) we drop κ from the notation.
(2) Let
Pκ (V)⊗ ⊂ P(V)⊗
be the full suboperad spanned by the κ-compact objects of P(V).
Remark 4.22. Since the Yoneda-embedding is monoidal and P(V)⊗ → Ass is compatible with small
colimits, Pκ (V)⊗ → Ass is a monoidal full subcategory of P(V)⊗ → Ass compatible with κ-small
colimits. The full subcategory P(V)κ of κ-compact objects of P(V) is the smallest full subcategory
of P(V) containing Pκ (V) and closed under retracts [17, Proposition 5.3.4.17.]. Thus P(V)κ is closed
under the tensor product of P(V) so that (P(V)κ )⊗ → Ass is monoidal ∞-category compatible with
κ-small colimits.
Remark 4.23. Since every object of P(V) is a small κ-filtered colimit of objects of Pκ (V), we
find that P(V)⊗ → Ass is a κ-compactly generated monoidal ∞-category. The monoidal ∞-category
Indκ (V)⊗ → Ass is a localization of P(V)⊗ relative to Ass. Consequently, a monoidal ∞-category is
κ-compactly generated if and only if it is a κ-accessible localization relative to Ass of presheaves on a
small monoidal ∞-category. Similarly, a bitensored ∞-category is κ-compactly generated if and only
if it is an enriched κ-accessible localization of presheaves on a small bitensored ∞-category.
Notation 4.24. Let κ be a small regular cardinal.
(1) Let V⊗ → Ass be a small monoidal ∞-category compatible with κ-small colimits, W⊗ → Ass a
small ∞-operad and M⊛ → V⊗ × W⊗ a small left κ-enriched ∞-category. Let
LEnr ⊂ V ×Indκ (V)⊗ PBEnv(M)LEnrκ ×PEnv(W)⊗ Pκ Env(W)
PBEnvκ (M)⊛ ⊗ ⊛ ⊗
be the full bitensored subcategory generated by M under κ-small colimits and the V, Pκ Env(W)-
biaction.
(2) Let V⊗ → Ass be a small ∞-operad, W⊗ → Ass a small monoidal ∞-category compatible with
κ-small colimits and M⊛ → V⊗ × W⊗ a small right κ-enriched ∞-category. Let
REnr ⊂ Pκ Env(V) ×PEnv(V)⊗ PBEnv(M)REnrκ ×Indκ (W)⊗ W
be the full bitensored subcategory generated by M under κ-small colimits and the Pκ Env(V), W-
biaction.
(3) Let V⊗ → Ass, W⊗ → Ass be small monoidal ∞-categories compatible with κ-small colimits
and M⊛ → V⊗ × W⊗ a small κ, κ-bi-enriched ∞-category. Let
BEnr ⊂ V ×Indκ (V)⊗ PBEnv(M)BEnrκ ×Indκ (W)⊗ W
be the full bitensored subcategory generated by M under κ-small colimits and the V, W-
biaction.
Remark 4.25. Since the embeddings V ⊂ Indκ (V) and Pκ Env(V) ⊂ PEnv(V) preserve κ-small col-
imits,
LEnr → V ×Pκ Env(W) , PBEnvκ (M)REnr → Pκ Env(V) ×W , PBEnvκ (M)BEnr → V ×W
PBEnvκ (M)⊛ ⊗ ⊗ ⊛ ⊗ ⊗ ⊛ ⊗ ⊗
are bitensored ∞-categories compatible with κ-small colimits, which are small because M is small and
the collection of all κ-small ∞-categories is small.
Notation 4.26. Let σ be the large strongly inaccessible cardinal corresponding to the small universe.
(1) Let V⊗ → Ass be a monoidal ∞-category compatible with small colimits, W⊗ → Ass an ∞-
operad and M⊛ → V⊗ × W⊗ a left quasi-enriched ∞-category. Let
PV (M)⊛ ∶= PBEnv(M)⊛ ∶= PBEnv
̂
LEnr σ (M)
⊛
→ V⊗ × PEnv(W)⊗ .
LEnr
(2) Let V → Ass be an ∞-operad, W → Ass a monoidal ∞-category compatible with small
⊗ ⊗
colimits and M⊛ → V⊗ × W⊗ a right quasi-enriched ∞-category. Let

PBEnv(M)⊛ ∶= PBEnv
̂
REnr σ (M)
⊛
→ PEnv(V)⊗ × W⊗ .
REnr
(3) Let V⊗ → Ass, W⊗ → Ass be monoidal ∞-categories compatible with small colimits and M⊛ →
V⊗ × W⊗ a bi-quasi-enriched ∞-category. Let
PV,W (M)⊛ ∶= PBEnv(M)⊛ ∶= PBEnv
BEnr
̂ σ (M)
⊛
→ V⊗ × W⊗ .
BEnr
Proposition 4.27. Let κ be a small regular cardinal.

(1) Let V⊗ → Ass be a small monoidal ∞-category compatible with κ-small colimits, W⊗ → Ass a
small ∞-operad, M⊛ → V⊗ ×W⊗ a small left κ-enriched ∞-category and N⊛ → V⊗ ×Pκ Env(W)⊗
a small bitensored ∞-category compatible with κ-small colimits. The functor
γN ∶ EnrFunV,Pκ Env(W) (PBEnvκ (M)LEnr , N) → EnrFunV,W (M, N)
admits a fully faithful left adjoint that lands in LinFunκV,Pκ Env(W) (PBEnvκ (M)LEnr , N) and
so the following induced functor is an equivalence:
LinFunκV,Pκ Env(W) (PBEnvκ (M)LEnr , N) → EnrFunV,W (M, N).
(2) Let V⊗ → Ass, W⊗ → Ass be small monoidal ∞-categories compatible with κ-small colimits,
M⊛ → V⊗ ×W⊗ a small κ, κ-bi-enriched ∞-category and N⊛ → V⊗ ×W⊗ a bitensored ∞-category
compatible with κ-small colimits. The functor
γN ∶ EnrFunV,W (PBEnvκ (M)BEnr , N) → EnrFunV,W (M, N)
admits a fully faithful left adjoint that lands in LinFunκV,W (PBEnvκ (M)BEnr , N) and so the
following induced functor is an equivalence:
LinFunκV,W (PBEnvκ (M)BEnr , N) → EnrFunV,W (M, N).
Proof. The proof of (1) and (2) are similar. We prove (1). Since PBEnvκ (M)LEnr is generated by
M under κ-small colimits and the V, Pκ Env(W)-biaction, the functor in the second part of (1) is
conservative. So it remains to prove the first part. By Notation 2.101 there is a κ-small colimits
preserving embedding of bitensored ∞-categories N⊛ → O⊛ ∶= Indκ (N)⊛ into a bitensored ∞-category
compatible with small colimits that lies over the monoidal embeddings V⊗ → Indκ (V)⊗ , Pκ Env(W)⊗ →
Indκ (Pκ Env(W))⊗ ≃ PEnv(W)⊗ . By Proposition 2.62 and Remark 4.25 the induced functors
ρ ∶ EnrFunIndκ (V),PEnv(W) (PBEnv(M)LEnrκ , O) → EnrFunV,Pκ Env(W) (PBEnvκ (M)LEnr , O),
γO ∶ EnrFunV,Pκ Env(W) (PBEnvκ (M)LEnr , O) → EnrFunV,W (M, O)
admit fully faithful left adjoints. Thus the left adjoint of γO factors as the left adjoint of γO ○ ρ, which
lands in LinFunLIndκ (V),PEnv(W) (PBEnv(M)LEnrκ , O) by Proposition 4.20, followed by ρ and so lands
in LinFunκV,Pκ Env(W) (PBEnvκ (M)LEnr, O) as PBEnvκ (M)LEnr is closed in PBEnv(M)LEnrκ under κ-
small colimits and the V, Pκ Env(W)-biaction. Since PBEnvκ (M)LEnr is generated by M under κ-small
colimits and bitensors and the linear embedding N⊛ → O⊛ preserves κ-small colimits, the left adjoint
of γO restricts to a left adjoint of γN that lands in LinFunV,Pκ Env(W) (PBEnvκ (M)LEnr, N).
Proposition 4.27 specializes to the following theorem:

Theorem 4.28. (1) Let V⊗ → Ass be a monoidal ∞-category compatible with small colimits,
W → Ass an ∞-operad, M⊛ → V⊗ × W⊗ a left quasi-enriched ∞-category and N⊛ → V⊗ ×
⊗
PEnv(W)⊗ a bitensored ∞-category compatible with small colimits. The induced functor is an
equivalence:
LinFunLV,PEnv(W) (PV (M), N) → EnrFunV,W (M, N).
(2) Let V⊗ → Ass, W⊗ → Ass be monoidal ∞-categories compatible with small colimits, M⊛ →
V⊗ ×W⊗ a bi-quasi-enriched ∞-category and N⊛ → V⊗ ×W⊗ a bitensored ∞-category compatible
with small colimits. The induced functor is an equivalence:
LinFunLV,W (PV,W (M), N) → EnrFunV,W (M, N).
Remark 4.29. Hinich [14, 6.2.2.], [15], [10, Notation 4.35., Theorem 5.1.] construct a different model
for the ∞-category of enriched presheaves in Hinich’s and Gepner-Haugseng’s models of enriched ∞-
categories, respectively. These constructions satisfy an analogous universal property in the framework
of these models and are compared by [13]. Consequently, comparison of these models with our model
[10], [18] identifies these constructions with the one of Theorem 4.28.
4.2. On the relationship between weak enrichment, pseudo enrichment and enrichment.
In the following we study the relationship between enriched, pseudo-enriched and weakly enriched
∞-categories.
Proposition 4.30. Let (V⊗ → Ass, S), (W⊗ → Ass, T) be small localization pairs.
(1) The functor
S−1 PEnv(V) ωBEnrW → V ωBEnrW ,
which takes pullback along the embedding V⊗ ⊂ S−1 PEnv(V)⊗ , restricts to an equivalence
S−1 PEnv(V) LEnrW → SV LEnrW .
(2) The functor
V ωBEnrT−1 PEnv(W) → V ωBEnrW ,
which takes pullback along the embedding W⊗ ⊂ T−1 PEnv(W)⊗ , restricts to an equivalence
V REnrT−1 PEnv(W) → V REnrT
W.
Proof. (2) is dual to (1). We prove (1). We first show that the functor S−1 PEnv(V) LEnrW → SV LEnrW
is conservative. Let F ∶ M⊛ → N⊛ be a S−1 PEnv(V), W-enriched functor between left enriched ∞-
categories whose pullback to V⊗ is an equivalence. Then F induces an equivalence on underlying
∞-categories and it is enough to see that F induces an equivalence on multi-morphism objects. Since
S−1 PEnv(V) is generated under small colimits by the tensor products of objects in the essential image
of j ∶ V ⊂ PEnv(V), it is enough to see that for every X, Y ∈ M and V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for
n, m ≥ 0 the following map is an equivalence:
PEnv(V)(j(V1 ) ⊗ ... ⊗ j(Vn ), MulMorM (X, W1 , ..., Wm , Y)) →
PEnv(V)(j(V1 ) ⊗ ... ⊗ j(Vn ), MulMorN (F(X), W1 , ..., Wm , F(Y))).
The latter map identifies with following map which is an equivalence by assumption:
MulM (j(V1 ), ..., j(Vn ), X, W1 , ..., Wm , Y) → MulN (j(V1 ), ..., j(Vn ), F(X), W1 , ..., Wm , F(Y)).
Let M⊛ → V⊗ × W⊗ be a totally small left S-enriched ∞-category. There is an embedding
M ⊂ PBEnv(M)⊛ S . Let M̄ → S PEnv(V) × W be the full weakly bi-enriched subcategory of
⊛ ⊛ −1 ⊗ ⊗
PBEnv(M)S → S PEnv(V) × W spanned by M. Since the latter is a presentably bitensored ∞-

⊛ −1 ⊗ ⊗
category, M̄⊛ → S−1 PEnv(V)⊗ × W⊗ is a left enriched ∞-category whose pullback to V⊗ × W⊗ is

M⊛ → V⊗ × W⊗ . We prove that for any small left enriched ∞-category N⊛ → S−1 PEnv(V)⊗ × W⊗ the
induced functor
αN ∶ EnrFunS−1 PEnv(V),W (M̄, N) → EnrFunV,W (M, N)
is an equivalence. The functor αN is the pullback of the functor
EnrFunS−1 PEnv(V),PEnv(W) (M̄, PBEnv(N)LEnr ) ÐÐÐÐÐÐÐÐ→ EnrFunV,PEnv(W)(M, PBEnv(N)LEnr ).

αPBEnv(N)LEnr
So it is enough to see that the latter functor is an equivalence. Since αPBEnv(N)LEnr is conservative,
it is an equivalence if it admits a fully faithful left adjoint. This follows from Proposition 2.62 and
Proposition 2.93 (2) because the left adjoint of the functor
AlgS−1 PEnv(V) (S−1 PEnv(V)) Ð→ AlgPEnv(V) (S−1 PEnv(V)) Ð

→ AlgV (S−1 PEnv(V))
L∗ κ
sends the map of ∞-operads τ ∶ V⊗ → PEnv(V)⊗ → S−1 PEnv(V)⊗ to the identity, where L ∶ PEnv(V)⊗ ⇆
S−1 PEnv(V)⊗ ∶ ι is the monoidal localization. Indeed, the left adjoint of κ sends τ to the monoidal
localization functor L ∶ PEnv(V)⊗ → S−1 PEnv(V)⊗ , which is sent by the left adjoint of L∗ given by ι∗
to L ○ ι ≃ id.
We complete the proof by showing that the functor S−1 PEnv(V) ωBEnrW → V ωBEnrW restricts to a
functor S−1 PEnv(V) LEnrW → V LEnrSW . If this is shown, the latter functor is an equivalence since it is
conservative and admits a fully faithful left adjoint as αN is an equivalence. First consider the case that
S is empty. In this case there is nothing to show and therefore the induced functor λ ∶ PEnv(V) LEnrW →
V ωBEnrW is an equivalence. For any left enriched ∞-category β ∶ M → S PEnv(V) × W let
⊛ −1 ⊗ ⊗
N → V × W be the pullback of β to V and M → PEnv(V) × W the pullback of β along the left

⊛ ⊗ ⊗ ⊗ ′⊛ ⊗ ⊗
adjoint monoidal functor L ∶ PEnv(V)⊗ → S−1 PEnv(V)⊗ . The weakly bi-enriched ∞-category M′⊛ →
PEnv(V)⊗ × W⊗ exhibits M as left enriched pulling back along a left adjoint. Because the pullback
of M′⊛ → PEnv(V)⊗ × W⊗ to V⊗ × W⊗ is N⊛ → V⊗ × W⊗ and λ is an equivalence, there is a canonical
PEnv(V), W-enriched equivalence N̄⊛ ≃ M′⊛ . The weakly bi-enriched ∞-category N⊛ → V⊗ × W⊗
exhibits N as left S-enriched in V because for every f ∈ S, X, Y ∈ N, W1 , ..., Wm ∈ W the induced map
MulN̄ (f, X, W1 , ..., Wm , Y) ≃ MulM′ (f, X, W1 , ..., Wm , Y) ≃ MulM (L(f), X, W1 , ..., Wm , Y)
is an equivalence.

Corollary 4.31. Let (V⊗ → Ass, S), (W⊗ → Ass, T) be small localization pairs. The functor
ρ ∶ S−1 PEnv(V) ωBEnrT−1 PEnv(W) → V ωBEnrW ,
which takes pullback along the embeddings V⊗ ⊂ S−1 PEnv(V)⊗ , W⊗ ⊂ T−1 PEnv(W)⊗ , restricts to an
equivalence
S−1 PEnv(V) BEnrT−1 PEnv(W) → V BEnrW .
S T
Proof. By Proposition 4.30 the induced functor

S−1 PEnv(V) LEnrT−1 PEnv(W) → SV LEnrT−1 PEnv(W)
is an equivalence, which restricts to an equivalence
S−1 PEnv(V) BEnrT−1 PEnv(W) → SV LEnrT−1 PEnv(W) ∩ V REnrT−1 PEnv(W) .
By Proposition 4.30 the induced functor
V REnrT−1 PEnv(W) → V RPEnrT
W
is an equivalence, which restricts to an equivalence

S
V LEnrT−1 PEnv(W) ∩ V REnrT−1 PEnv(W) → SV LEnrW ∩ V RPEnrT S T
W = V BPEnrW .

Proposition 4.30 gives the following corollaries:
Corollary 4.32. Let V⊗ → Ass, W⊗ → Ass be small ∞-operads.
(1) The functor
PEnv(V) LEnrW → V ωBEnrW
taking pullback along the embedding V⊗ ⊂ PEnv(V)⊗ is an equivalence.

(2) The functor
V REnrPEnv(W) → V ωBEnrW
taking pullback along the embedding W⊗ ⊂ PEnv(W)⊗ is an equivalence, which restricts to an

equivalence
V BEnrPEnv(W) → V ωLEnrW .
(3) The functor

PEnv(V) BEnrPEnv(W) → V ωBEnrW
taking pullback along the embeddings V⊗ ⊂ PEnv(V)⊗ , W⊗ ⊂ PEnv(W)⊗ is an equivalence.
(4) Let V⊗ → Ass be a small monoidal ∞-category. The functor
P(V) LEnrW → V LPEnrW
taking pullback along the monoidal embedding V⊗ ⊂ P(V)⊗ is an equivalence.
(5) Let W⊗ → Ass be a small monoidal ∞-category. The functor
V REnrP(W) → V RPEnrW
taking pullback along the monoidal embedding W⊗ ⊂ P(W)⊗ is an equivalence.
(6) Let V⊗ → Ass, W⊗ → Ass be small monoidal ∞-categories. The functor
P(V) BEnrP(W) → V BPEnrW
taking pullback along the monoidal embeddings V⊗ ⊂ P(V)⊗ , W⊗ ⊂ P(W)⊗ is an equivalence.

Remark 2.55 and Corollary 4.32 give the following corollary:
Corollary 4.33. (1) The functors S LEnr∅ → Cat∞ , ∅ REnrS → Cat∞ , S BEnrS → Cat∞ taking
pullback along the embedding of ∞-operads ∅⊗ ⊂ S× are equivalences.
(2) The functors ∗ LPEnr∅ → Cat∞ , ∅ RPEnr∗ → Cat∞ , ∗ BPEnr∗ → Cat∞ taking pullback along the
embedding of ∞-operads ∅⊗ ⊂ Ass are equivalences.
Corollary 4.34. Let κ, τ be small regular cardinals.
(1) Let V⊗ → Ass be a small monoidal ∞-category compatible with κ-small colimits and W⊗ → Ass
a small ∞-operad. The functor
Indκ (V) LEnrW → κV LEnrW
taking pullback along the monoidal embedding V⊗ ⊂ Indκ (V)⊗ is an equivalence.
(2) Let W⊗ → Ass be a small monoidal ∞-category compatible with τ -small colimits and V⊗ → Ass
a small ∞-operad. The functor
V REnrIndτ (W) → V REnrτW
taking pullback along the monoidal embedding W⊗ ⊂ Indτ (W)⊗ is an equivalence.
(3) Let V⊗ → Ass, W⊗ → Ass be small monoidal ∞-categories compatible with κ-small colimits,
τ -small colimits, respectively. The functor
Indκ (V) BEnrIndτ (W) → κV BEnrτW
taking pullback along the monoidal embeddings V⊗ ⊂ Indκ (V)⊗ , W⊗ ⊂ Indτ (W)⊗ is an equiva-
lence.
Corollary 4.35. (1) Let V⊗ → Ass be a presentably monoidal ∞-category, W⊗ → Ass a small
∞-operad and M⊛ → V⊗ × W⊗ , N⊛ → V⊗ × W⊗ left enriched ∞-categories such that M, N are
small. The ∞-category EnrFunV,W (M, N) is small.
(2) Let V⊗ → Ass be a small ∞-operad, W⊗ → Ass a presentably monoidal ∞-category and M⊛ →
V⊗ ×W⊗ , N⊛ → V⊗ ×W⊗ right enriched ∞-categories such that M, N are small. The ∞-category
EnrFunV,W (M, N) is small.
(3) Let M⊛ → V⊗ × W⊗ , N⊛ → V⊗ × W⊗ be ∞-categories bi-enriched in presentably monoidal
∞-categories such that M, N are small. The ∞-category EnrFunV,W (M, N) is small.
For the next corollary we use Notation 4.21:
Corollary 4.36. Let κ, τ be small regular cardinals and V⊗ → Ass, W⊗ → Ass be small ∞-operads.
(1) The functor
PEnvκ (V) LEnrW → V ωBEnrW
κ
taking pullback along the embedding V⊗ ⊂ PEnvκ (V)⊗ is an equivalence.

(2) The functor
V REnrPEnvτ (W) → V ωBEnrW
τ
taking pullback along the embedding W⊗ ⊂ PEnvτ (W)⊗ is an equivalence.

(3) The functor
PEnvκ (V) BEnrPEnvτ (W) → V ωBEnrW
κ τ
taking pullback along the embeddings V⊗ ⊂ PEnvκ (V)⊗ , W⊗ ⊂ PEnvτ (W)⊗ is an equivalence.
Proof. The functor PEnv(V) LEnrW → V ωBEnrW taking pullback along the embedding V⊗ ⊂ PEnv(V)⊗ ,
which is an equivalence by Corollary 4.32 (1), factors as PEnv(V) LEnrW → κPEnvκ (V) LEnrW → V ωBEnrW .
By Corollary 4.34 (1) the first functor in the composition is an equivalence since the embedding
PEnvκ (V) ⊂ PEnv(V) induces an equivalence Indκ (PEnvκ (V)) ≃ PEnv(V). The proof of (2) and (3)
are similar and follow from Corollary 4.32 and 4.34 (2) and (3), respectively.

Next we compare different sorts of tensored ∞-categories (Corollary 4.38).
Lemma 4.37. Let κ be a small regular cardinal and V⊗ → V′⊗ , W⊗ → W′⊗ embeddings of ∞-operads
into a monoidal ∞-category compatible with κ-small colimits such that V′ , W′ are generated under
κ-small colimits by tensor products of objects in the essential image.
(1) A left κ-enriched ∞-category N⊛ → V′⊗ × W⊗ is left tensored compatible with κ-small colimits
if and only if the pullback V⊗ ×V′⊗ N⊛ → V⊗ × W⊗ admits κ-small conical colimits and left
tensors.
(2) A left κ-enriched ∞-category N⊛ → V′⊗ × W⊗ admits right tensors if and only if the pullback
V⊗ ×V′⊗ N⊛ → V⊗ × W⊗ admits right tensors.
(3) A right κ-enriched ∞-category N⊛ → V⊗ × W′⊗ is right tensored compatible with κ-small
colimits if and only if the pullback N⊛ ×W′⊗ W⊗ → V⊗ × W⊗ admits κ-small conical colimits
and right tensors.
(4) A right κ-enriched ∞-category N⊛ → V⊗ × W′⊗ admits left tensors if and only if the pullback
N⊛ ×W′⊗ W⊗ → V⊗ × W⊗ admits left tensors.
(5) A κ, κ-bi-enriched ∞-category N⊛ → V′⊗ × W′⊗ is bitensored compatible with κ-small colimits
if and only if the pullback V⊗ ×V′⊗ N⊛ ×W′⊗ W⊗ → V⊗ × W⊗ admits left and right tensors and
κ-small conical colimits.
Proof. (5) follows immediately from (1) - (4). The proof of (3) is similar to (1). The proof of (4)
is similar to (2). We prove (1) and (2). We start with proving (2). Let W ∈ W, X ∈ N and X ⊗ W
the right tensor for φ ∶ N⊛ → V′⊗ × W⊗ . Then there is an object in MulN (X, W; X ⊗ W) and by left
pseudo-enrichedness of φ it is enough to prove that for every V′ ∈ V′ , W1 , ..., Wm ∈ W for m ≥ 0 the
following induced map is an equivalence:
ρ ∶ MulN (V′ , X ⊗ W, W1 , ..., Wm ; Y) → MulN (V′ , X, W, W1 , ..., Wm ; Y).
By left κ-enrichedness of φ we can moreover assume that V′ ≃ V1 ⊗ ... ⊗ Vn for V1 , ..., Vn ∈ V and n ≥ 0
since V′ is generated by tensor products of V under κ-small colimits. Again by left pseudo-enrichedness
of φ for this choice of V′ the map ρ identifies with the map
MulN (V1 , ...Vn , X ⊗ W, W1 , ..., Wm ; Y) → MulN (V1 , ..., Vn , X, W, W1 , ..., Wm ; Y),
which is an equivalence by universal property of the right tensor.
We continue with proving (1). The only-if direction is clear. We prove the if-direction. Assume
that N⊛ → V′⊗ × W⊗ exhibits N as left κ-enriched and the pullback V⊗ ×V′⊗ N⊛ → V⊗ × W⊗ admits
left tensors and κ-small conical colimits. We like to see that N⊛ → V′⊗ × W⊗ admits left tensors
and κ-small conical colimits. Since V⊗ ×V′⊗ N⊛ → V⊗ × W⊗ admits left tensors, by iterately taking
left tensors for every V1 , ..., Vn ∈ V, X ∈ N for n ≥ 0 the functor MulN (V1 , ..., Vn , X; −) ∶ N → S
is corepresentable. This implies that for every V ∈ V′ , X ∈ N the functor MulN (V, X; −) ∶ N → S
is corepresentable, too. This holds since by assumption V′ is generated under κ-small colimits by
tensor products of objects of V, the ∞-category N admits κ-small colimits, and for every X ∈ V′ the
functor V′op → Fun(N, S), V′ ↦ MulN (V′ , X; −) preserves κ-small limits. The latter holds as for every
Y ∈ N, V ∈ V′ the functor MulN (−, X; Y) ∶ V′op → S preserves κ-small limits by left κ-enrichedness.
Let V ⊗ X be the object corepresenting the functor MulN (V, X; −) ∶ N → S. We obtain a functor
⊗ ∶ V′ ×N → N that sends V1 ⊗...⊗Vn ∈ V′ for V1 , ..., Vn ∈ V for n ≥ 0 to (V1 ⊗(−))○...○(Vn⊗(−)) ∶ N → N.
Moreover we obtain a natural transformation α ∶ (−, X) → (−) ⊗ X of functors V′ → N⊛ that sends any
U′ ∈ V′ to the universal multi-morphism U′ , X → U′ ⊗ X in N⊛ . We prove that for every U′ ∈ V′ , X ∈ N
the universal morphism U′ , X → U′ ⊗ X in N⊛ exhibits U′ ⊗ X as the left tensor of U′ and X. For every
X ∈ N the natural transformation α ∶ (−, X) → (−) ⊗ X induces for every V′ ∈ V′ , W1 , ..., Wm ∈ W for
m ≥ 0 a natural transformation
β ∶ MulN (V′ , (−) ⊗ X, W1 , ..., Wm ; Y) → MulN (V′ , −, X, W1 , ..., Wm ; Y)
of functors V′op × V′op → S. By left pseudo-enrichedness of N⊛ → V′⊗ × W⊗ the universal morphism
U′ , X → U′ ⊗ X in N⊛ exhibits U′ ⊗ X as the left tensor of U′ and X if β is an equivalence. By left κ-
enrichedness of N⊛ → V′⊗ × W⊗ we can assume that V′ ≃ V1 ⊗ ... ⊗ Vk for some V1 , ..., Vk ∈ V and k ≥ 0.
Moreover by left κ-enrichedness of N⊛ → V′⊗ × W⊗ the target of β preserves κ-small limits. We prove
next that also the source of β preserves κ-small limits. The functor (−) ⊗ X ∶ V′ → N preserves κ-small
colimits because the functor MulN (−, X; Y) ∶ V′op → S preserves κ-small limits by left κ-enrichedness
of N⊛ → V′⊗ × W⊗ . Therefore it is enough to see that the functor MulN (V′ , −, W1 , ..., Wm ; Y) ∶
Nop → S preserves κ-small limits. By left pseudo-enrichedness of N⊛ → V′⊗ × W⊗ the canonical map
V1 , ..., Vk → V′ in V′⊗ identifies the functor MulN (V′ , −, W1 , ..., Wm ; Y) ∶ Nop → S with the functor
MulN (V1 , ..., Vk , −, W1 , ..., Wm ; Y) ∶ Nop → S, which preserves κ-small limits because the pullback
V⊗ ×V′⊗ N⊛ → V⊗ × W⊗ admits κ-small conical colimits. Hence source and target of β preserve κ-small
limits. This implies that it is enough to prove that for every U1 , ..., Un ∈ V for n ≥ 0 the map βU1 ⊗...⊗Un
is an equivalence. The map βU1 ⊗...⊗Un factors as equivalences
MulN (V′ , (U1 ⊗ ... ⊗ Un ) ⊗ X, W1 , ..., Wm ; Y) ≃ MulN (V1 , ..., Vk , (U1 ⊗ ... ⊗ Un ) ⊗ X, W1 , ..., Wm ; Y)
≃ MulN (V1 , ..., Vk , U1 , (U2 ⊗ ... ⊗ Un ) ⊗ X, W1 , ..., Wm ; Y)

≃ MulN (V1 , ..., Vk , U1 , U2 , (U3 ⊗ ... ⊗ Un ) ⊗ X, W1 , ..., Wm ; Y) ≃ ...
≃ MulN (V1 , ..., Vk , U1 , ..., Un , X, W1 , ..., Wm ; Y) ≃ MulN (V′ , U1 ⊗ ... ⊗ Un , X, W1 , ..., Wm ; Y).
It remains to prove that for every V′ ∈ V′ the functor V′ ⊗(−) ∶ N → N preserves κ-small colimits. By
assumption this holds for every V′ ∈ V and so also for every V′ ≃ V1 ⊗...⊗Vn for V1 , ..., Vn ∈ V for n ≥ 0.
The full subcategory of V′ spanned by all V′ such that the functor V′ ⊗ (−) ∶ N → N preserves κ-small
colimits is closed under κ-small colimits because the functor V′ → Fun(N, N) ∶ V′ ↦ V′ ⊗ (−) ∶ N → N
preserves κ-small colimits as the functor MulN (−, X; Y) ∶ V′op → S preserves κ-small limits by left
κ-enrichedness of N⊛ → V′⊗ × W⊗ . So the result follows from the fact that V′ is generated under
κ-small colimits by tensor products of objects of V.

For the next corollary we use Notation 4.21:
Corollary 4.38. Let κ be a small regular cardinal and V⊗ → Ass, W⊗ → Ass be small ∞-operads.
(1) The functor
PEnvκ (V) LEnrW → V ωBEnrW
κ
taking pullback along the embedding V⊗ ⊂ PEnvκ (V)⊗ restricts to an equivalence between left
tensored ∞-categories compatible with κ-small colimits and weakly bi-enriched ∞-categories
that admit left tensors and κ-small conical colimits.
(2) The functor
V REnrPEnvκ (W) → V ωBEnrW
κ
taking pullback along the embedding W⊗ ⊂ PEnvκ (W)⊗ restricts to an equivalence between right
tensored ∞-categories compatible with κ-small colimits and weakly bi-enriched ∞-categories
that admit right tensors and κ-small conical colimits.
(3) The functor
PEnvκ (V) BEnrPEnvτ (W) → V ωBEnrW
κ κ
taking pullback along the embeddings V⊗ ⊂ PEnvκ (V)⊗ , W⊗ ⊂ PEnvκ (W)⊗ restricts to an
equivalence between bitensored ∞-categories compatible with κ-small colimits and weakly bi-
enriched ∞-categories that admit left and right tensors and κ-small conical colimits.
4.3. Presentability of enriched presheaves. Next we prove that the ∞-category of presheaves of
Notation 4.26 enriched in a presentably monoidal ∞-category is presentable (Corollary 4.42).
Notation 4.39. Let κ, τ be small regular cardinals and M⊛ → V⊗ × W⊗ a weakly bi-enriched ∞-
category.
(1) Let M⊛κ → (V ) × W be the pullback of M → V × W along the embedding of ∞-operads
κ ⊗ ⊗ ⊛ ⊗ ⊗
κ ⊗ ⊗
(V ) ⊂ V .
(2) Let M⊛κ,τ → (V ) × (W ) be the pullback of M⊛ → V⊗ × W⊗ along the embeddings of
κ ⊗ τ ⊗
∞-operads (V ) ⊂ V , (W ) ⊂ W⊗ .
κ ⊗ ⊗ τ ⊗
Proposition 4.40. Let κ, τ be small regular cardinals.

(1) Let V⊗ → Ass be a κ-compactly generated monoidal ∞-category, W⊗ → Ass a small ∞-operad
and M⊛ → V⊗ × W⊗ a small left enriched ∞-category. By Corollary 4.34 (1) the Vκ , W-
enriched embedding M⊛ ⊛
κ ⊂ PBEnv(Mκ )LEnrκ extends to a V, W-enriched embedding M
⊛
⊂
PBEnv(Mκ )LEnrκ . For every bitensored ∞-category compatible with small colimits N →
⊛ ⊛
V⊗ × PEnv(W)⊗ the following functor is an equivalence:

LinFunLV,PEnv(W) (PBEnv(Mκ )LEnrκ , N) → EnrFunV,W (M, N).
(2) Let V⊗ → Ass, W⊗ → Ass be κ-compactly generated, τ -compactly generated monoidal ∞-

categories, respectively, and M⊛ → V⊗ ×W⊗ a small bi-enriched ∞-category. By Corollary 4.34
(3) the Vκ , Wτ -enriched embedding M⊛ ⊛
κ,τ ⊂ PBEnv(Mκ,τ )BEnrκ,τ extends to a V, W-enriched
embedding M⊛ ⊂ PBEnv(Mκ,τ )⊛ BEnrκ,τ . For every bitensored ∞-category compatible with small
colimits N⊛ → V⊗ × W⊗ the following functor is an equivalence:
BEnrκ,τ , N) → EnrFunV,W (M, N).
LinFunLV,W (PBEnv(Mκ,τ )⊛
Proof. By Corollary 4.34 (1) the functor EnrFunV,W (M, N) → EnrFunVκ ,W (Mκ , Nκ ) is an equiva-
lence. So we need to see that LinFunLV,PEnv(W) (PBEnv(Mκ )LEnrκ , N) → EnrFunVκ ,W (Mκ , Nκ ) is an
equivalence. Let L ∶ PEnv(Vκ )⊗ → P(Vκ )⊗ → Indκ (Vκ )⊗ ≃ V⊗ be the left adjoint relative to Ass of
the canonical embedding. The latter functor factors as
LinFunLV,PEnv(W) (PBEnv(Mκ )LEnrκ , N) → LinFunLPEnv(Vκ ),PEnv(W) (PBEnv(Mκ ), L∗ (N))
→ EnrFunVκ ,W (Mκ , Nκ ).
The first functor is an equivalence by the description of local equivalences. The second functor is an
equivalence by Corollary 4.32 (1).
(2): By Corollary 4.34 (3) the functor EnrFunV,W (M, N) → EnrFunVκ ,Wτ (Mκ,τ , Nκ,τ ) is an equiv-
alence. So it is enough to see that LinFunLV,W (PBEnv(Mκ,τ )BEnrκ,τ , N) → EnrFunVκ ,Wτ (Mκ,τ , Nκ,τ )
is an equivalence. The latter functor factors as
LinFunLV,W (PBEnv(Mκ,τ )BEnrκ,τ , N) → LinFunLPEnv(Vκ ),PEnv(Wτ ) (PBEnv(Mκ,τ ), (L, L)∗ (N))
→ EnrFunVκ ,Wτ (Mκ,τ , Nκ,τ ).
The first functor is an equivalence by the description of local equivalences. The second functor is
an equivalence by Corollary 4.32 (3).

Theorem 4.28 and Proposition 4.40 give the following description of the ∞-category of presheaves
enriched in a presentably monoidal ∞-category.
and M⊛ → V⊗ × W⊗ a small left enriched ∞-category. The enriched embedding M⊛ κ ⊂ M
⊛
induces an equivalence
PBEnv(Mκ )⊛ ⊛
LEnrκ ≃ PBEnv(M)LEnr .
(2) Let V⊗ → Ass, W⊗ → Ass be κ-compactly generated, τ -compactly generated monoidal ∞-

categories, respectively, and M⊛ → V⊗ × W⊗ a small bi-enriched ∞-category. The enriched
embedding M⊛ ⊛
κ,τ ⊂ M induces an equivalence
PBEnv(Mκ,τ )⊛ ⊛
BEnrκ,τ ≃ PBEnv(M)BEnr .

and M⊛ → V⊗ ×W⊗ a small left enriched ∞-category. Then PBEnv(M)⊛ LEnr → V ×PEnv(W)
⊗ ⊗
is a κ-compactly generated bitensored ∞-category.

(2) Let V⊗ → Ass be a κ-compactly generated monoidal ∞-category, W⊗ → Ass a τ -compactly
generated monoidal ∞-category and M⊛ → V⊗ × W⊗ a small bi-enriched ∞-category. Then
PBEnv(M)⊛ BEnr → V ×W is a λ-compactly generated bitensored ∞-category for every regular
⊗ ⊗
cardinal λ ≥ κ, τ.
Proposition 4.43. Let M⊛ → V⊗ × W⊗ be a totally small weakly bi-enriched ∞-category.
(1) Let M̄⊛ → PEnv(V)⊗ × W⊗ be the unique extension to a left enriched ∞-category. Then M̄⊛ ⊂
PBEnv(M)⊛ induces an equivalence of ∞-categories bitensored over PEnv(V), PEnv(W):
PBEnv(M̄)⊛ ⊛
LEnr ≃ PBEnv(M) .
(2) Let M̄⊛ → PEnv(V)⊗ × PEnv(W)⊗ the unique extension to a bi-enriched ∞-category. Then
M̄⊛ ⊂ PBEnv(M)⊛ induces an equivalence of ∞-categories bitensored over PEnv(V), PEnv(W):
PBEnv(M̄)⊛ ⊛
BEnr ≃ PBEnv(M) .
Proof. (1): Let N⊛ → PEnv(V)⊗ × PEnv(W)⊗ be a locally small bitensored ∞-category compatible
with small colimits. The PEnv(V), PEnv(W)-enriched embedding M̄⊛ ⊂ PBEnv(M̄)⊛LEnr induces an
equivalence
LinFunLPEnv(V),PEnv(W) (PBEnv(M̄)LEnr , N) → EnrFunPEnv(V),W (M̄, N) ≃ EnrFunV,W (M, N)
by Theorem 4.28 (1), where the last equivalence is by Corollary 4.32 (1). We apply Corollary 4.42.
(2): Let N⊛ → PEnv(V)⊗ ×PEnv(W)⊗ be a locally small bitensored ∞-category compatible with small
colimits. The PEnv(V), PEnv(W)-enriched embedding M̄⊛ ⊂ PBEnv(M̄)⊛ BEnr induces an equivalence
LinFunLPEnv(V),PEnv(W) (PBEnv(M̄)BEnr , N) → EnrFunPEnv(V),PEnv(W)(M̄, N) ≃ EnrFunV,W (M, N)
by Theorem 4.28 (3), where the last equivalence is by Corollary 4.32 (3). We apply Corollary 4.42.

Corollary 4.41, Corollary 2.109 and Corollary 4.36 give the following corollary:
Corollary 4.44. Let κ, τ be small regular cardinals and M⊛ → V⊗ × W⊗ a totally small weakly bi-
(1) Let M̄⊛ → PEnv(V)⊗ × W⊗ be the unique extension to a left enriched ∞-category. Then M̄⊛
κ ⊂
PBEnv(M)⊛ induces an equivalence of ∞-categories bitensored over PEnv(V), PEnv(W):
PBEnv(M̄κ )⊛ ⊛
LEnrκ ≃ PBEnv(M) .
(2) Let M̄⊛ → PEnv(V)⊗ × PEnv(W)⊗ be the unique extension to a bi-enriched ∞-category. Then
κ,τ ⊂ PBEnv(M) induces an equivalence of ∞-categories bitensored over PEnv(V), PEnv(W):
M̄⊛ ⊛
PBEnv(M̄κ,τ )⊛ ⊛
BEnrκ,τ ≃ PBEnv(M) .
4.4. Atomicity of representable enriched presheaves.

Definition 4.45. Let V⊗ → Ass, W⊗ → Ass be a presentably monoidal ∞-category and M⊛ → V⊗ ×W⊗
a locally small bi-enriched ∞-category. An object X ∈ M is atomic if the V, W-enriched functor
MorM (X, −) ∶ M⊛ → (V ⊗ W)⊛
is linear and preserves small colimits.
In this subsection we prove the following theorem:
Theorem 4.46. Let M⊛ → V⊗ × W⊗ be a totally small weakly bi-enriched ∞-category and (V⊗ →
Ass, S), (W⊗ → Ass, T) small localization pairs. The essential image of the functor
M ⊂ PBEnv(M) Ð
→ PBEnv(M)S,T
L
consists of atomic objects.

Lemma 4.47. Let V⊗ → Ass, W⊗ → Ass be small monoidal ∞-categories, M⊛ → P(V × W)⊗ a biten-
sored ∞-category compatible with small colimits and N⊛ → P(V)⊗ the underlying left tensored ∞-
category. Let X, Y ∈ M, V ∈ V, W ∈ W and F ∈ P(V). The canonical morphism
α ∶ F ⊗ LMorN (X ⊗ W, Y) → LMorN (X ⊗ W, F ⊗ Y)
in P(V) evaluated at V is equivalent to the canonical morphism
β ∶ F ⊗ LMorM (X, Y) → LMorM (X, F ⊗ Y)
in P(V × W) evaluated at V, W.
Proof. Let V ∈ V. Evaluating α at V gives the canonical map
ψ ∶ colimV′ ,V′′ ∈V,V→V′ ⊗V′′ F(V′ ) × M(V′′ ⊗ X ⊗ W, Y) → M(V ⊗ X ⊗ W, F ⊗ Y)
in P(V). Evaluating β at (V, W) gives the canonical map
colimV′ ,V′′ ∈V,W′ ,W′′ ∈W,V→V′⊗V′′ ,W→W′′ ⊗W′ F(V′ )×W(W′ , 1W )×M(V′′ ⊗X⊗W′′ , Y) → M(V⊗X⊗W, F⊗Y)
in P(V × W). The latter map factors as the following map in P(V × W) and so identifies with ψ ∶
colimV′ ,V′′ ∈V,V→V′ ⊗V′′ (F(V′ ) × colimW′ ,W′′ ∈W,W→W′′ ⊗W′ W(W′ , 1W ) × M(V′′ ⊗ X ⊗ W′′ , Y)) →
´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹¸ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹¶
(1W ⊗M(V′′ ⊗X⊗(−),Y))(W)
M(V ⊗ X ⊗ W, F ⊗ Y).

Proposition 4.48. Let M⊛ → V⊗ be a weakly left enriched ∞-category, V ∈ V, X, Y ∈ M and α ∈
MulM (V, X, Y) a multi-morphism that exhibits V as the left morphism object LMorM (X, Y) ∈ V of
X, Y. For every V1 , ..., Vn ∈ V for n ≥ 0 the induced morphism
V1 ⊗ ... ⊗ Vn ⊗ V ⊗ X → V1 ⊗ ... ⊗ Vn ⊗ Y
in LEnv(M) exhibits V1 ⊗ ... ⊗ Vn ⊗ V as the left morphism object LMorLEnv(M) (X, V1 ⊗ ... ⊗ Vn ⊗ Y) ∈
Env(V) of X, V1 ⊗ ... ⊗ Vn ⊗ Y.
In particular, the morphism object of X, V1 ⊗ ... ⊗ Vn ⊗ Y in LEnv(M) exists and the following
canonical morphism in Env(V) is an equivalence:
V1 ⊗ ... ⊗ Vn ⊗ LMorLEnv(M) (X, Y) → LMorLEnv(M) (X, V1 ⊗ ... ⊗ Vn ⊗ Y).
In particular, the following canonical morphism in PEnv(V) is an equivalence:
V1 ⊗ ... ⊗ Vn ⊗ LMorPLEnv(M) (X, Y) → LMorPLEnv(M) (X, V1 ⊗ ... ⊗ Vn ⊗ Y).
Proof. We like to prove that for any V′1 , ...V′ℓ ∈ V for ℓ ≥ 0 the following induced map is an equivalence:
(11) Env(V)(V′1 ⊗...⊗V′ℓ , V1 ⊗...⊗Vn ⊗LMorM (X, Y)) → LEnv(M)(V′1 ⊗...⊗V′ℓ ⊗X, V1 ⊗...⊗Vn ⊗Y).
For any [n], [m] ∈ Ass let
Ass([m], [n]) ⊂ Assmin,max ([m], [n]) ⊂ Assmin ([m], [n])
be the subsets of order preserving maps preserving the minimum and maximum, preserving the
minimum, respectively. The map Ass([ℓ], [n + 1]) = ∆([n + 1], [ℓ]) → ∆([n], [ℓ]) = Ass([ℓ], [n])
restricting along the order preserving embedding [n] ≅ {0, ..., n} ⊂ [n + 1] induces a bijection ξ ∶
Assmin,max ([ℓ], [n + 1]) → Assmin ([ℓ], [n]). The map (11) canonically covers ξ. Thus we have to prove
that for every order preserving map ϕ ∶ [n + 1] → [ℓ] preserving the minimum the map induced by
(11) on the fiber over ϕ is an equivalence. This map identifies with the map
n
∏ Env(V)(Vϕ(i−1)+1 ⊗ ... ⊗ Vϕ(i) , Vi ) × Env(V)(Vϕ(n)+1 ⊗ ... ⊗ Vℓ , LMorLEnv(M) (X, Y)) →
′ ′ ′ ′
i=1
n
∏ Env(V)(Vϕ(i−1)+1 ⊗ ... ⊗ Vϕ(i) , Vi ) × LEnv(M)(Vϕ(n)+1 ⊗ ... ⊗ Vℓ ⊗ X, Y).
′ ′ ′ ′
i=1
induced by the identity of ∏ni=1 Env(V)(Vϕ(i−1)+1
′
⊗ ... ⊗ Vϕ(i)
′
, Vi ) and the canonical map
′
Env(V)(Vϕ(n)+1 ⊗ ... ⊗ Vℓ′ , LMorLEnv(M) (X, Y)) → LEnv(M)(Vϕ(n)+1
′
⊗ ... ⊗ Vℓ′ ⊗ X, Y).
The latter map identifies with the map
′
MulV (Vϕ(n)+1 , ..., Vℓ′ , LMorLEnv(M) (X, Y)) → MulM (Vϕ(n)+1
′
, ..., Vℓ′ , X, Y),
which is an equivalence. Consequently, the morphism object of X, V1 ⊗ ... ⊗ Vn ⊗ Y in LEnv(M) exists
and the following canonical morphism in Env(V) is an equivalence:
V1 ⊗ ... ⊗ Vn ⊗ LMorM (X, Y) → LMorLEnv(M) (X, V1 ⊗ ... ⊗ Vn ⊗ Y).
By Proposition 3.13 the morphism object of X, Y in LEnv(M) exists if the morphism object of X, Y
in M exists, and the canonical morphism LMorM (X, Y) → LMorLEnv(M) (X, Y) in Env(V) is an equiv-
alence. This proves the second part. The third part follows from the fact that the linear embedding
LEnv(M)⊛ ⊂ PLEnv(M)⊛ lying over the monoidal embedding Env(V)⊗ ⊂ PEnv(V)⊗ preserves left
morphism objects by Proposition 3.13.

Proof of Theorem 4.46. We first reduce to the case that S, T are empty. Consider the PEnv(V), PEnv(W)-
enriched adjunction
(−) ⊗ X ⊗ (−) ∶ (PEnv(V) ⊗ PEnv(W))⊛ ⇄ PBEnv(M)⊛ ∶ LMorPBEnv(M) (X, −).
The left adjoint preserves local equivalences for the localizations
PBEnv(M) → PBEnv(M)S,T , PEnv(V) ⊗ PEnv(W) → S−1 PEnv(V) ⊗ T−1 PEnv(W).
If the statement holds for S, T empty, then the right adjoint is PEnv(V), PEnv(W)-linear and preserves
small colimits. This implies that also the right adjoint preserves local equivalences. Consequently,
in this case the latter PEnv(V), PEnv(W)-enriched adjunction induces a S−1 PEnv(V), T−1 PEnv(W)-
enriched adjunction
S,T ∶ LMorPBEnv(M)S,T (L(X), −),
(−) ⊗ L(X) ⊗ (−) ∶ (S−1 PEnv(V) ⊗ T−1 PEnv(W))⊛ ⇄ PBEnv(M)⊛
in which the right adjoint preserves small colimits and is S−1 PEnv(V), T−1 PEnv(W)-linear. So it
suffices to prove the case that S, T are empty.
The PEnv(V), PEnv(W)-enriched functor
LMorPBEnv(M) (X, −) ∶ PBEnv(M)⊛ → P(Env(V) × Env(W))⊛
preserves small colimits since for any V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W the composition
PBEnv(M) ÐÐÐÐÐÐÐÐÐÐÐ→ P(Env(V) × Env(W)) ÐÐÐÐÐÐÐÐÐÐÐÐÐ→ S

LMorPBEnv(M) (X,−) evV1 ⊗...⊗Vn ,W1 ⊗...⊗Wm
identifies with the map PBEnv(M)(V⊗X⊗W, −) ∶ PBEnv(M) → S evaluating at V⊗X⊗W ∈ BEnv(M).

It remains to see that the PEnv(V), PEnv(W)-enriched functor
is PEnv(V), PEnv(W)-linear. We first reduce to the case that W⊗ = ∅⊗ . Let N⊛ ∶= REnv(M)⊛ →
V⊗ × Env(W)⊗ . By Remark 2.96 there is a canonical equivalence PBEnv(M)⊛ ≃ PLEnv(N)⊛ of ∞-
categories bitensored over PEnv(V), PEnv(W). In particular, the underlying left PEnv(V)-tensored
∞-category of PBEnv(M)⊛ → PEnv(V)⊗ × PEnv(W)⊗ is PLEnv(N)⊛ → PEnv(V)⊗ . By Lemma 4.47
the PEnv(V), PEnv(W)-enriched functor
is PEnv(V), PEnv(W)-linear if for every W ∈ PEnv(W) the left PEnv(V)-enriched functor

LMorPLEnv(N) (X ⊗ W, −) ∶ PLEnv(N)⊛ → PEnv(V)⊛
is left PEnv(V)-linear. So have reduced to the case that W⊗ = ∅⊗ and need to see that for ev-
ery X ∈ M the left PEnv(V)-enriched functor LMorPLEnv(M) (X, −) ∶ PLEnv(M)⊛ → PEnv(V)⊛ is
left PEnv(V)-linear. We reduce next to the case that M⊛ → V⊗ is a left enriched ∞-category.
Let M̄⊛ ⊂ PLEnv(M)⊛ → PEnv(V)⊗ be the full left PEnv(V)-enriched ∞-category spanned by M.
Corollary 4.44 guarantees that the embedding M̄⊛ ω ⊂ PLEnv(M) lying over the embedding of ∞-
⊛
ω ⊗ ⊗
operads (PEnv(V) ) ⊂ PEnv(V) uniquely extends to a left adjoint PEnv(V)-linear equivalence
PLEnv(M̄ω )⊛ LEnrω ≃ PLEnv(M) . Thus the left PEnv(V)-enriched functor LMorPLEnv(M) (X, −) ∶
⊛
PLEnv(M)⊛ → PEnv(V)⊛ identifies with the left PEnv(V)-enriched functor LMorPLEnv(M̄)LEnrω (X, −) ∶
PLEnv(M̄ω )⊛ LEnr → PEnv(V) ≃ Indω (PEnv(V) ) . So replacing PEnv(V) by V it is enough to show
⊛ ω ⊛ ω
that for every totally small left enriched ∞-category M → V and X ∈ M the left V-enriched functor
⊛ ⊗
LMorPLEnv(M)LEnrω (X, −) ∶ PLEnv(M)⊛ LEnrω → Indω (V) is left V-linear. By the first part of the proof
⊛
this holds if the left PEnv(V)-enriched functor LMorPLEnv(M) (X, −) ∶ PLEnv(M)⊛ → PEnv(V)⊛ is
left PEnv(V)-linear. This way we reduced to the case that M⊛ → V⊗ is a left enriched ∞-category.
So it remains to see that for every X ∈ M, Y ∈ PLEnv(M) and V ∈ PEnv(V) the following canonical
morphism in PEnv(V) is an equivalence:
V ⊗ LMorPLEnv(M) (X, Y) → LMorPLEnv(M) (X, V ⊗ Y).
Since PLEnv(M) → PEnv(V)⊗ is a left tensored ∞-category compatible with small colimits and
⊛
PLEnv(M) is generated under small colimits by objects of the form V1 ⊗ ... ⊗ Vn ⊗ Y for V1 , ..., Vn ∈ V
for n ≥ 0 and Y ∈ M, it is enough to see that the canonical morphism
σ ∶ V1 ⊗...⊗Vn ⊗LMorPLEnv(M) (X, W1 ⊗...⊗Wm ⊗Y) → LMorPLEnv(M) (X, V1 ⊗...⊗Vn ⊗W1 ⊗...⊗Wm ⊗Y)
is an equivalence. The composition
V1 ⊗...⊗Vn ⊗W1 ⊗...⊗Wm ⊗LMorPLEnv(M) (X, Y) → V1 ⊗...⊗Vn ⊗LMorPLEnv(M) (X, W1 ⊗...⊗Wm ⊗Y)
Ð
→ LMorPLEnv(M) (X, V1 ⊗ ... ⊗ Vn ⊗ W1 ⊗ ... ⊗ Wm ⊗ Y)
σ
is the canonical morphism. Consequently, it suffices to see that for every V1 , ..., Vn ∈ V for n ≥ 0 and
Y ∈ M the canonical morphism
V1 ⊗ ... ⊗ Vn ⊗ LMorPLEnv(M) (X, Y) → LMorPLEnv(M) (X, V1 ⊗ ... ⊗ Vn ⊗ Y)
is an equivalence. This holds by Proposition 4.48.

Corollary 4.49. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category. The functor
ξ ∶ Mop → EnrFunV,W (M, PEnv(V) ⊗ PEnv(W))
sending X ∈ M to LMorM̄ (X, −) corresponding to the functor ΓM of Construction 3.6 is an embedding.
Proof. By construction 3.6 the functor ξ factors as an embedding
α ∶ Mop → EnrFunPEnv(V),PEnv(W) (PBEnv(M), PEnv(V) ⊗ PEnv(W))
sending X ∈ M to LMorPBEnv(M) (X, −) followed by restriction
EnrFunPEnv(V),PEnv(W) (PBEnv(M), PEnv(V) ⊗ PEnv(W)) → EnrFunV,W (M, PEnv(V) ⊗ PEnv(W)).
Proposition 4.46 for S, T empty guarantees that α lands in the full subcategory
EnrFunLPEnv(V),PEnv(W)(PBEnv(M), PEnv(V) ⊗ PEnv(W)).
By Corollary 2.109 following restriction is an equivalence:

EnrFunLPEnv(V),PEnv(W) (PBEnv(M), PEnv(V) ⊗ PEnv(W)) → EnrFunV,W (M, PEnv(V) ⊗ PEnv(W)).

Corollary 4.50. Let (V⊗ → Ass, S), (W⊗ → Ass, T) be small localization pairs, M⊛ → V⊗ × W⊗ a to-
tally small S, T-bi-enriched ∞-category, N⊛ → S−1 PEnv(V)⊗ ×T−1 PEnv(W)⊗ a locally small bitensored
∞-category compatible with small colimits and φ ∶ PBEnv(M)⊛ S,T → N a small colimits preserving
⊛
−1 −1
S PEnv(V), T PEnv(W)-linear functor.
(1) Then φ is an embedding if the restriction φ∣M⊛ ∶ M⊛ → N⊛ is an embedding and for every
X ∈ M the image φ(X) is atomic.
(2) Then φ is an equivalence if and only if the restriction φ∣M⊛ ∶ M⊛ → N⊛ is an embedding, for
every X ∈ M the image φ(X) is atomic and N is generated under small colimits and left and
right tensors by the essential image of φ∣M⊛ .
Proof. (1) We like to see that for every X, Y ∈ PBEnv(M)S,T the induced morphism
αX,Y ∶ MorPBEnv(M)S,T (X, Y) → MorN (φ(X), φ(Y))
in S−1 PEnv(V) ⊗ T−1 PEnv(W) is an equivalence. By assumption αX,Y is an equivalence if X, Y ∈ M.
We prove first that for every Y ∈ PBEnv(M)S,T the full subcategory ΘY of PBEnv(M)S,T spanned
by those X such that αX,Y is an equivalence, is closed under small colimits and the biaction. This
implies the claim if we have shown that M ⊂ ΘY since PBEnv(M)S,T is generated by M under small
colimits and the biaction as a consequence of Proposition 2.93. The full subcategory ΘY is closed
under small colimits since φ preserves small colimits and for every Z ∈ N the functor MorN (−, Z) ∶
Nop → PEnv(V) ⊗ PEnv(W) preserves small limits as N⊛ → PEnv(V)⊗ × PEnv(W)⊗ is a bitensored
∞-category compatible with small colimits, and similar for PBEnv(M)S,T . The full subcategory
ΘY is closed under the biaction since for every V ∈ S−1 PEnv(V), W ∈ T−1 PEnv(W) the induced
morphism MorPBEnv(M)S,T (V⊗X⊗W, Y) → MorN (φ(V⊗X⊗W), φ(Y)) ≃ MorN (V⊗φ(X)⊗W, φ(Y))
identifies with the induced morphism S−1 PEnv(V) ⊗ T−1 PEnv(W)(V ⊗ W, MorPBEnv(M)S,T (X, Y)) →
S−1 PEnv(V) ⊗ T−1 PEnv(W)(V ⊗ W, MorN (φ(X), φ(Y))). It remains to see that M ⊂ ΘY . Because
PBEnv(M)S,T is generated by M under small colimits and the biaction, it is enough to prove that
for every X ∈ M the full subcategory of PBEnv(M)S,T spanned by those Y such that αX,Y is an
equivalence, is closed under small colimits and the biaction. This follows from Proposition 4.46.
(2) follows immediately from (1), Remark 4.17, Proposition 4.46 and Proposition 2.93.

Corollary 4.51. Let V⊗ → Ass be an ∞-operad and A an associative algebra in V. Let B(A)⊛ ⊂
RModA (V)⊛ be the full weakly left enriched subcategory spanned by A.
Corollary 4.52. Let V⊗ → Ass be a presentably monoidal ∞-category and A an associative alge-
bra in V. The left V-enriched embedding B(A)⊛ ⊂ RModA (V)⊛ induces a left V-linear equivalence
PBEnv(B(A))⊛ ⊛
LEnr ⊂ RModA (V) .
Corollary 4.53. We apply Corollary 4.50 and observe that RModA (V) is generated by A under left
tensors and that A is atomic in RModA (V).
Proposition 4.54. Let (V⊗ → Ass, S), (W⊗ → Ass, T) be small localization pairs and M⊛ → V⊗ × W⊗
a totally small weakly bi-enriched ∞-category. Let M′⊛ ⊂ PBEnv(M)⊛S,T be the full weakly bi-enriched
subcategory spanned by V, W and the essential image of the functor M ⊂ PBEnv(M) Ð

→ PBEnv(M)S,T .
L
(1) The weakly bi-enriched ∞-category M′⊛ → V⊗ × W⊗ exhibits M′ as S, T-bi-enriched in V, W.

(2) For every X, Y ∈ M the induced morphism MorPBEnv(M) (X, Y) → MorPBEnv(M′ ) (L(X), L(Y))
in PEnv(V) ⊗ PEnv(W) is a local equivalence for the localization PEnv(V) ⊗ PEnv(W) →
S−1 PEnv(V) ⊗ T−1 PEnv(W).
(3) For every weakly bi-enriched ∞-category N⊛ → V⊗ × W⊗ that exhibits N as S, T-bi-enriched in
V, W the induced V, W-enriched functor M⊛ → M′⊛ induces an equivalence
ρN ∶ EnrFunV,W (M′ , N) → EnrFunV,W (M, N).
Proof. (1): Let M̂′⊛ ⊂ PBEnv(M)⊛ be the full weakly bi-enriched subcategory spanned M′ . Then
M′⊛ → V⊗ × W⊗ is the pullback of the bi-enriched ∞-category M̂′⊛ → PEnv(V)⊗ × PEnv(W)⊗ .
Let M̄′⊛ ⊂ PBEnv(M′ )⊛ be the full weakly bi-enriched subcategory spanned M′ . Then M′⊛ →
V⊗ × W⊗ is the pullback of the bi-enriched ∞-category M̄′⊛ → PEnv(V)⊗ × PEnv(W)⊗ . By Proposi-
tion 4.30 the induced PEnv(V), PEnv(W)-linear functor PBEnv(M)⊛ → PBEnv(M′ )⊛ restricts to a
PEnv(V), PEnv(W)-enriched equivalence M̂′⊛ → M̄′⊛ since the pullback of the latter enriched functor
to V⊗ , W⊗ is the identity of M′⊛ . Hence for every X, Y ∈ M, V ∈ PEnv(V), W ∈ PEnv(W) the following
map is an equivalence:
(12) PBEnv(M)(V ⊗ L(X) ⊗ W, L(Y)) → PBEnv(M′ )(V ⊗ L(X) ⊗ W, L(Y)).
Since L∣M ∶ M → M′ is essentially surjective, we need to see that for every X, Y ∈ M, f ∈ S, g ∈ T the
map PBEnv(M′ )(f ⊗ L(X) ⊗ g, L(Y)) is an equivalence. By equivalence (12) the latter map identifies
with the map PBEnv(M)(f ⊗L(X)⊗g, L(Y)), which is an equivalence by Lemma 4.19 since L(Y) ∈ M′ .
(2): For every X, Y ∈ M the induced morphism MorPBEnv(M) (X, Y) → MorPBEnv(M′ ) (L(X), L(Y))
in PEnv(V) ⊗ PEnv(W) identifies with the following morphism via equivalence (12):
MorPBEnv(M) (X, Y) → MorPBEnv(M) (X, L(Y)) ≃ MorPBEnv(M) (L(X), L(Y))
≃ MorPBEnv(M′ ) (L(X), L(Y)).
By Theorem 4.46 the enriched functor MorPBEnv(M) (X, −) ∶ PBEnv(M)⊛ → (PEnv(V) ⊗ PEnv(W))⊛
is linear and preserves small colimits and therefore sends (generating) local equivalences for the local-
ization PBEnv(M) → PBEnv(M)S,T to local equivalences for the localization PEnv(V) ⊗ PEnv(W) →
S−1 PEnv(V) ⊗ T−1 PEnv(W).
(3): The functor ρN is the pullback of the functor ρPBEnv(N)S,T . So it is enough to see that the
latter functor is an equivalence. Consider the commutative triangle:
LinFunLS−1 PEnv(V),T−1 PEnv(W) (PBEnv(M)S,T , PBEnv(N)S,T )

❨❨❨❨❨❨
❨❨❨❨❨❨ α
❨❨❨❨❨❨
❨❨❨❨❨❨
❨❨❨,
EnrFunV,W (M, PBEnv(N)S,T )
ρPBEnv(N)S,T ❡❡❡❡❡❡❡❡
❡2
❡ ❡❡❡ ❡❡
❡
❡❡❡❡❡❡
❡❡❡❡❡❡
EnrFunV,W (M′ , PBEnv(N)S,T )
By Proposition 4.20 the functor α is an equivalence. So it is enough to see that the vertical
functor is an equivalence. By Proposition 4.20 it is enough to prove that the induced left ad-
joint S−1 PEnv(V), T−1 PEnv(W)-linear functor θ ∶ PBEnv(M′ )⊛
S,T → PBEnv(M)S,T is an equivalence.
⊛
The ∞-category PBEnv(M) is generated by M under small colimits and the PEnv(V), PEnv(W)-
biaction. Hence the localization PBEnv(M)S,T is generated by M′ under small colimits and the
S−1 PEnv(V), T−1 PEnv(W)-biaction. Therefore by Corollary 4.50 the functor θ is an equivalence if for
every X ∈ M the S−1 PEnv(V), T−1 PEnv(W)-enriched functor
MorPBEnv(M)S,T (L(X), −) ∶ PBEnv(M)⊛
S,T → (S PEnv(V) ⊗ T PEnv(W))
−1 −1 ⊛
is linear and preserves small colimits. This holds by Theorem 4.46.

Corollary 4.55. Let (V⊗ → Ass, S), (W⊗ → Ass, T) be small localization pairs. The embedding
S T
V BEnrW ⊂ V ωBEnrW
admits a left adjoint. A V, W-enriched functor φ ∶ M⊛ → N⊛ is a local equivalence if and only if

● the underlying functor M → N is essentially surjective,
● for every objects X, Y ∈ M the induced morphism MorPBEnv(M) (X, Y) → MorPBEnv(N) (φ(X), φ(Y))
in PEnv(V) ⊗ PEnv(W) is a local equivalence for the localization
PEnv(V) ⊗ PEnv(W) → S−1 PEnv(V) ⊗ T−1 PEnv(W).
Corollary 4.56. (1) Let (V⊗ → Ass, S), (W⊗ → Ass, T) be small localization pairs. The ∞-
category V BEnrT
S
W is compactly generated.
(2) For every presentably monoidal ∞-categories V⊗ → Ass, W⊗ → Ass the ∞-category V BEnrW
is compactly generated.
Proof. (1): By Proposition 2.57 the ∞-category V ωBEnrW is compactly generated. By Proposition
4.54 the embedding SV BEnrT W ⊂ V ωBEnrW , which preserves filtered colimits, admits a left adjoint. So
the claim follows.
(2): Let κ, τ be small regular cardinals such that V⊗ → Ass is κ-compactly generated and W⊗ → Ass
is τ -compactly generated. By Corollary 4.34 (3) the canonical functor V BEnrW → κVκ BEnrτWτ is an
equivalence. So the claim follows from (1) and Example 4.9.

4.5. Transfer of enrichment via scalar extension. In this subsection we describe and study the
process of transfering (weak) enrichment along maps of ∞-operads (Proposition 4.61).
Notation 4.57. Let M⊛ → U⊗ × V⊗ , N⊛ → V⊗ × W⊗ , O⊛ → W⊗ × Q⊗ be presentably bitensored

∞-categories. We write (M ⊗V N ⊗W O)⊛ → U⊗ × Q⊗ for the relative tensor product of [16, Definition
4.4.2.10.].
Lemma 4.58. Let M⊛ → V⊗ be a presentably right tensored ∞-category, N⊛ → V⊗ × W⊗ a presentably

bitensored ∞-category and O⊛ → W⊗ a presentably left tensored ∞-category. The ∞-category M ⊗V
N⊗W O is generated under small colimits by the essential image of the functor M×N×O → M⊗V N⊗W O.
Proof. By [16, Example 4.7.2.7.] there is a canonical equivalence
M ⊗V N ⊗W O ≃ colim[n]∈Ass (M ⊗ V⊗n ⊗ N ⊗ W⊗n ⊗ O).
So every object of M ⊗V N ⊗W O lies in the essential image of the canonical small colimits preserving
functor λn ∶ M⊗V⊗n ⊗N⊗W⊗n ⊗O → M⊗V N⊗W O for some [n] ∈ Ass. The map [0] ≃ {0} ⊂ [n] induces
a map M ⊗ V⊗n ⊗ N ⊗ W⊗n ⊗ O → M ⊗ N ⊗ O over M ⊗V N ⊗W O so that the functor λ0 ∶ M ⊗ N ⊗ O →
M ⊗V N ⊗W O is essentially surjective. There is a universal functor α ∶ M × N × O → M ⊗ N ⊗ O that
preserves small colimits component-wise. We will prove that M ⊗ N ⊗ O is generated under small
colimits by the essential image of α. This will imply that M ⊗V N ⊗W O is generated under small
colimits by the essential image of λ0 ○ α.
Since M, N, O are presentable, there are small ∞-categories A, B, C and localizations P(A) ⇆
M, P(B) ⇆ N, P(C) ⇆ O. By [16, Proposition 4.8.1.17.] we obtain an induced localization P(A) ⊗
P(B) ⊗ P(C) ⇆ M ⊗ N ⊗ O. Since there is a commutative square
P(A) × P(B) × P(C) / P(A) ⊗ P(B) ⊗ P(C)

α
M×N×O / M ⊗ N ⊗ O,
α

it is enough to see that P(A) ⊗ P(B) ⊗ P(C) is generated under small colimits by the essential image
of the functor α ∶ P(A) × P(B) × P(C) → P(A) ⊗ P(B) ⊗ P(C). By [16, Proposition 4.8.1.17.] there is a
canonical equivalence P(A) ⊗ P(B) ⊗ P(C) ≃ P(A × B × C) and the left adjoint functor α identifies with
the canonical left adjoint functor P(A) × P(B) × P(C) → P(A × B × C) whose restriction to A × B × C
is the Yoneda-embedding. So the result follows from the fact that P(A × B × C) is generated under
small colimits by A × B × C [17, Corollary 5.1.5.8.].

Corollary 4.59. Let M⊛ → V⊗ ×W⊗ be a presentably bitensored ∞-category and V⊗ → V′⊗ , W⊗ → W′⊗
small colimits preserving monoidal functors of presentably monoidal ∞-categories. Then V′ ⊗V M ⊗W
W′ is generated under small colimits and the V′ , W′ -biaction by the essential image of the functor
M → V′ ⊗V M ⊗W W′ .
Proof. This follows immediately from Lemma 4.58 since the functor M → V′ ⊗V M ⊗W W′ refines to a
V, W-linear functor.

Corollary 4.60. Let M⊛ → V⊗ × W⊗ , N⊛ → V⊗ × W⊗ be presentably bitensored ∞-categories, G ∶
N⊛ → M⊛ a V, W-linear functor whose underlying functor preserves small colimits and is monadic,
that admits a V, W-linear left adjoint F. Let O⊛ → V⊗ be a presentably right tensored ∞-category
and P⊛ → W⊗ a presentably left tensored ∞-category. The functor O ⊗V N ⊗W P → O ⊗V M ⊗W P is
monadic.
Proof. The V, W-enriched adjunction F ∶ M⊛ ⇄ N⊛ ∶ G, where both adjoints preserve small colimits
and are V, W-linear, gives rise to an adjunction O⊗V F⊗W P ∶ O⊗V M⊗W P ⇄ O⊗V N⊗W P ∶ O⊗V G⊗W P,
where both adjoints preserve small colimits. So by the monadicity theorem [16, Theorem 4.7.3.5.] the
right adjoint is monadic if it is conservative. We observe that the right adjoint is conservative if
O ⊗V N ⊗W P is generated under small colimits by the essential image of the left adjoint. We will prove
that O ⊗V N ⊗W P is generated under small colimits by the essential image of the left adjoint. Let
Q ⊂ O ⊗V N ⊗W P be a full subcategory containing the essential image of O ⊗V F ⊗W P and closed under
small colimits. For every X ∈ O, Y ∈ P let QX,Y ⊂ N the full subcategory spanned by all Z ∈ N whose
image under the canonical small colimits preserving functor {X} × N × {Y} → O × N × P → O ⊗ N ⊗ P →
O ⊗V N ⊗W P belongs to Q. Then QX,Y is closed in N under small colimits and contains the essential
image of F ∶ M → N. Since G is monadic, N is generated under small colimits by the essential image
of F [16, Theorem 4.7.3.5.]. Hence QX,Y = N so that Q contains the essential image of the functor
O × N × P → O ⊗ N ⊗ P → O ⊗V N ⊗W P. Hence Q = O ⊗V N ⊗W P by Lemma 4.58.

Theorem 4.61. The forgetful functor γ ∶ ωBEnr → Op∞ × Op∞ is a cocartesian fibration. Let
ψ ∶ M⊛ → N⊛ be an enriched functor lying over maps of ∞-operads α ∶ V⊗ → V′⊗ , α′ ∶ W⊗ → W′⊗ . The
following conditions are equivalent:
(1) The map ψ is γ-cocartesian.
(2) The induced linear functor BEnv(M)⊛ → BEnv(N)⊛ exhibits BEnv(N) as
Env(V′ ) ⊗Env(V) BEnv(M) ⊗Env(W) Env(W′ ).
(3) The functor M → N is essentially surjective and the induced linear functor PBEnv(M)⊛ →
PBEnv(N)⊛ exhibits PBEnv(N) as
PEnv(V′ ) ⊗PEnv(V) PBEnv(M) ⊗PEnv(W) PEnv(W′ ).
(4) The functor M → N is essentially surjective and for every X, Y ∈ M the induced morphism
(α, β)! (ΓM (X, Y)) → ΓN (ψ(X), ψ(Y))
in P(BEnv(V ) × BEnv(W′ )) is an equivalence.
′
Proof. We prove that (1) implies (2): Let ψ ∶ M⊛ → N⊛ be an enriched functor lying over maps of
∞-operads α ∶ V⊗ → V′⊗ , β ∶ W⊗ → W′⊗ . For every bitensored ∞-category Q⊛ → U⊗ × T ⊗ the square
BMod(BEnv(N), Q) / BMod(BEnv(M), Q)
Mon(Env(V′ ), U) × Mon(Env(W′ ), T) / Mon(Env(V), U) × Mon(Env(W), T)

identifies with the commutative square

(13) ωBEnr(N, Q) / ωBEnr(M, Q)
Op∞ (V′ , U) × Op∞ (W′ , T) / Op (V, U) × Op (W, T).

∞ ∞
Next we show that (3) implies (1). Let Q⊛ → U⊗ × T ⊗ be a weakly bi-enriched ∞-category. We
like to see that square (13) is a pullback square. If the functor M → N is essentially surjective, square
(13) is a pullback square if square (13) for Q⊛ replaced by PBEnv(Q)⊛ is a pullback square. But this
square (after rotation) identifies with the commutative square
ccBMod(PBEnv(N), PBEnv(Q)) / LMon(PEnv(V′ ), PEnv(U)) × LMon(PEnv(W′ ), PEnv(T))
/ LMon(PEnv(V), PEnv(U)) × LMon(PEnv(W), PEnv(T)).

ccBMod(PBEnv(M), PBEnv(Q))
Next we prove that (3) implies (4). By Example 2.51 for every X ∈ M there is a unique left adjoint
PEnv(V), PEnv(W)-linear functor P(Env(V) × Env(W))⊛ → PBEnv(M)⊛ sending the tensor unit to
X, which by Corollary 4.46 admits a PEnv(V), PEnv(W)-linear, small colimits preserving right adjoint
sending Z to LMorPBEnv(M) (X, Z). Applying the functor PEnv(V′ ) ⊗PEnv(V) (−) ⊗PEnv(W) PEnv(W′ )
the induced left adjoint PEnv(V′ ), PEnv(W′ )-linear functor
P(Env(V′ ) × Env(W′ ))⊛ → (PEnv(V′ ) ⊗PEnv(V) PBEnv(M) ⊗PEnv(W) PEnv(W′ ))⊛
is right adjoint to the PEnv(V′ ), PEnv(W′ )-linear, small colimits preserving functor
LMorPBEnv(N) (ψ(X), −).
Consequently, by uniqueness of right adjoints for every Y ∈ M there is a canonical equivalence
(α, β)! (LMorPBEnv(M) (X, Y)) ≃ LMorPBEnv(N) (ψ(X), ψ(Y)).
We will complete the proof by showing that γ is a cocartesian fibration whose cocartesian morphisms
satisfy (3) and so conditions (1), (2) and (4). This will imply that condition (3) follows from (2) and
that condition (1) follows from (4): the enriched functors satisfying conditions (2), (4), respectively,
satisfy the (2) out of (3)-property. By the existence of γ-cocartesian lifts (that satify conditions
(1)-(4)) we are reduced to show that every V, W-enriched functor ψ ∶ M⊛ → N⊛ satisfying (2), (4),
respectively, is an equivalence. The first fact follows from the fact that ψ is the pullback of the functor
BEnv(M)⊛ → BEnv(N)⊛ . If a V, W-enriched functor ψ ∶ M⊛ → N⊛ satisfies (4), the induced functor

M → N on underlying ∞-categories is essentially surjective and for every X, Y ∈ M the morphism
LMorPBEnv(M) (X, Y) → LMorPBEnv(N) (ψ(X), ψ(Y)) is an equivalence. So for every V1 , ..., Vn ∈ V for
n ≥ 0 the following map is an equivalence:
PEnv(V)(V1 ⊗ ... ⊗ Vn , LMorPBEnv(M) (X, Y)) → PEnv(V)(V1 ⊗ ... ⊗ Vn , LMorPBEnv(N) (ψ(X), ψ(Y))).
The latter map identifies with the map MulM (V1 , ..., Vn , X, Y) → MulN (V1 ...Vn , ψ(X), ψ(Y)). So if
(4) holds, ψ is an equivalence.
Hence it remains to prove that γ is a cocartesian fibration whose cocartesian morphisms satisfy
(3). Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category and α ∶ V⊗ → V′⊗ , β ∶ W⊗ → W′⊗
be maps of ∞-operads. Let (α, β)! (M)⊛ → V′⊗ × W′⊗ be the full weakly bi-enriched subcategory
of the presentably bitensored ∞-category (PEnv(V′ ) ⊗PEnv(V) PBEnv(M) ⊗PEnv(W) PEnv(W′ ))⊛ →
(PEnv(V′ ) ⊗ PEnv(W′ ))⊗ spanned by the essential image of the canonical functor
τ ∶ M ⊂ PBEnv(M) → PEnv(V′ ) ⊗PEnv(V) PBEnv(M) ⊗PEnv(W) PEnv(W′ )
and let ψ ∶ M⊛ → (α, β)! (M)⊛ be the canonical enriched functor. We will prove that the embedding
(α, β)! (M)⊛ ⊂ (PEnv(V′ ) ⊗PEnv(V) PBEnv(M) ⊗PEnv(W) PEnv(W′ ))⊛ induces an equivalence:
κ ∶ PBEnv((α, β)! (M))⊛ ≃ (PEnv(V′ ) ⊗PEnv(V) PBEnv(M) ⊗PEnv(W) PEnv(W′ ))⊛ .
This will imply that the induced left adjoint linear functor
PBEnv(M)⊛ → PBEnv((α, β)! (M))⊛ ≃ (PEnv(V′ ) ⊗PEnv(V) PBEnv(M) ⊗PEnv(W) PEnv(W′ ))⊛
is the canonical one so that ψ satisfies condition (3). To see that the functor κ is fully faithful, it is
enough to see that for every X ∈ M the PEnv(V′ ), PEnv(W′ )-enriched functor
LMorPEnv(V′ )⊗PEnv(V) PBEnv(M)⊗PEnv(W) PEnv(W′ ) (τ (X), −) ∶
(PEnv(V′ ) ⊗PEnv(V) PBEnv(M) ⊗PEnv(W) PEnv(W′ ))⊛ → (PEnv(V′ ) ⊗ PEnv(W′ ))⊛
is PEnv(V′ ), PEnv(W′ )-linear and preserves small colimits. The PEnv(V), PEnv(W)-linear adjunction
(−) ⊗ X ⊗ (−) ∶ (PEnv(V) ⊗ PEnv(W))⊛ ⇄ PBEnv(M)⊛ ∶ LMorPBEnv(M) (X, −),
where the right adjoint is PEnv(V), PEnv(W)-linear and preserves small colimits by Corollary 4.46,
is sent by the Cat∞ -linear functor
PEnv(V′ ) ⊗PEnv(V) (−) ⊗PEnv(W) PEnv(W′ ) ∶ PEnv(V) BModPEnv(W) → PEnv(V′ ) BModPEnv(W)′
to a PEnv(V′ ), PEnv(W′ )-linear adjunction
(−)⊗τ (X)⊗(−) ∶ (PEnv(V′ )⊗PEnv(W′ ))⊛ ⇄ (PEnv(V′ )⊗PEnv(V) PBEnv(M)⊗PEnv(W) PEnv(W′ ))⊛ ∶
PEnv(V′ ) ⊗PEnv(V) LMorPBEnv(M) (X, −) ⊗PEnv(W) PEnv(W′ ),
where the right adjoint is PEnv(V′ ), PEnv(W′ )-linear and preserves small colimits. Hence the latter
right adjoint identifies with the PEnv(V′ ), PEnv(W′ )-enriched functor
LMorPEnv(V′ )⊗PEnv(V) PBEnv(M)⊗PEnv(W) PEnv(W′ ) (τ (X), −),
which therefore becomes linear and preserves small colimits. Moreover κ is essentially surjective since
the ∞-category PEnv(V′ )⊗PEnv(V) PBEnv(M)⊗PEnv(W) PEnv(W′ ) is generated by the essential image
of τ under small colimits and the PEnv(V′ ), PEnv(W′ )-biaction according to Corollary 4.59.

Corollary 4.62. Let W⊗ → Ass be an ∞-operad and ψ ∶ M⊛ → N⊛ a right W-enriched functor lying
over a left adjoint monoidal functor between presentably monoidal ∞-categories α ∶ V⊗ → V′⊗ . If ψ is
γ-cocartesian, the functor PV,W (M)⊛ → PV′ ,W (N)⊛ exhibits PV′ ,W (N) as V′ ⊗V PV,W (M).
Corollary 4.63. Let ψ ∶ M⊛ → N⊛ be an enriched functor lying over left adjoint monoidal functors
between presentably monoidal ∞-categories α ∶ V⊗ → V′⊗ , α′ ∶ W⊗ → W′⊗ . If ψ is γ-cocartesian, the
functor PV,W (M)⊛ → PV′ ,W′ (N)⊛ exhibits PV′ ,W′ (N) as V′ ⊗V PV,W (M) ⊗W W′ .
Corollary 4.64. Let φ ∶ V⊗ → V′⊗ , ψ ∶ W⊗ → W′⊗ be maps of small ∞-operads. The functor (φ, ψ)∗ ∶
V′ ωBEnrW′ → V ωBEnrW admits a left adjoint (φ, ψ)! ∶ V ωBEnrW → V′ ωBEnrW′ .
Proposition 4.65. Let φ ∶ V⊗ → V′⊗ , ψ ∶ W⊗ → W′⊗ be maps of small ∞-operads and κ, τ small
regular cardinals.
(1) The induced functor (φ, ψ)! ∶ V ωLEnrW → V′ ωLEnrW′ of Proposition 4.61 restricts to functors
κ
V LPEnrW → V′ LPEnrκW′ ,
V REnrW → V′ RPEnrW′ ,
τ τ
V BPEnrW → V′ BPEnrW′
κ τ κ τ
if φ is a monoidal functor preserving κ-small colimits for the first functor, if ψ is a monoidal
functor preserving τ -small colimits for the second functor, if both φ and ψ are monoidal func-
tors and φ preserves κ-small colimits and ψ preserves τ -small colimits for the third functor.
(2) The induced functor (φ, ψ)! ∶ V ωLEnrW → V′ ωLEnrW′ of Proposition 4.61 restricts to functors
V LEnrW → V′ LEnrW′ ,
V REnrW → V′ REnrW′ ,
V BEnrW → V′ BEnrW′ ,
respectively, if ψ, φ, both φ and ψ admit a left adjoint relative to Ass, respectively.
Proof. (1) follows from Proposition 4.61 (4) in view of the fact that the induced left adjoint functor
φ! ∶ P(V) → P(V′ ) restricts to a functor Indκ (V) → Indκ (V′ ) and similar for ψ. The latter holds since
Indκ (V) is generated by V as full subcategory of P(V) under small κ-filtered colimits [17, Corollary
5.3.5.4.].
We prove (2) for the first functor of (2), the second case of (2) is dual and the third case of (2)
follows from the first and second one. If ψ admits a left adjoint α relative to Ass, also Env(α) is left
adjoint to Env(φ). So there is an adjunction
Fun(Env(α)op , PEnv(V′ )) ∶ Fun(Env(W)op , PEnv(V′ )) ⇄
Fun(Env(W′ )op , PEnv(V′ )) ∶ Fun(Env(ψ)op , PEnv(V′ )).
Thus the functor (Env(φ) × Env(ψ))! ∶ P(Env(V) × Env(W)) → P(Env(V′ ) × Env(W′ )) is equivalent
to the functor
Fun(Env(W)op , PEnv(V)) ÐÐÐÐÐÐÐÐÐÐ→ Fun(Env(W)op , PEnv(V′ ))
Fun(Env(W)op ,φ! )
ÐÐÐÐÐÐÐÐÐÐÐÐÐÐ→ Fun(Env(W′ )op , PEnv(V′ ))

Fun(Env(α)op ,PEnv(V′ ))
since the right adjoints are equivalent. The latter functor restricts to a functor
Fun(Env(W)op , V) ÐÐÐÐÐÐÐÐÐÐ→ Fun(Env(W)op , V′ ) ÐÐÐÐÐÐÐÐÐÐ→ Fun(Env(W′ )op , V′ ).

Fun(Env(W)op ,φ) Fun(Env(α)op ,V′ )

Proposition 4.61 gives the following corollary:
Corollary 4.66. Let W⊗ → Ass be a small ∞-operad and φ ∶ V⊗ → V′⊗ an embeddings of small ∞-
operads. The induced functor φ! ∶ V LEnrW → V′ LEnrW is fully faithful and the essential image precisely
consists of the bi-enriched ∞-categories whose left morphism objects lie in the essential image of φ.
Proof. By Proposition 4.61 there is an induced adjunction φ! ∶ V ωBEnrW ⇄ V′ ωBEnrW ∶ φ∗ . By

Proposition 4.61 (4) the left adjoint sends V BEnrW to the full subcategory Ξ of V′ ωBEnrW spanned
by the bi-enriched ∞-categories whose left multi-morphism objects lie in the essential image of φ.
We first prove that the right adjoint φ∗ sends Ξ to V BEnrW . Let M⊛ → V′⊗ × W⊗ be a bi-
enriched ∞-category that belongs to Ξ. Then for every X, Y ∈ M, W1 , ..., Wm ∈ W for m ≥ 0
the left multi-morphism object LMulMorM (X, W1 , ..., Wm ; Y) belongs to the essential image of φ
and so corresponds to an object of V denoted by φ−1 (LMulMorM (X, W1 , ..., Wm ; Y)). The univer-
sal morphism LMulMorM (X, W1 , ..., Wm ; Y), X, W1 , ..., Wm → Y in M⊛ corresponds to a morphism
φ−1 (LMulMorM (X, W1 , ..., Wm ; Y)), X, W1 , ..., Wm → Y in φ∗ (M)⊛ . For every V1 , ..., Vn ∈ V for n ≥ 0
the induced map
MulV (V1 , ..., Vn ; φ−1 (LMulMorM (X, W1 , ..., Wm ; Y))) → Mulφ∗ (M) (V1 , ..., Vn , X, W1 , ..., Wm ; Y)
identifies with the map
MulV (V1 , ..., Vn ; φ−1 (LMulMorM (X, W1 , ..., Wm ; Y))) ≃
MulV′ (φ(V1 ), ..., φ(Vn ); LMulMorM (X, W1 , ..., Wm ; Y)) ≃ MulM (φ(V1 ), ..., φ(Vn ), X, W1 , ..., Wm ; Y)
≃ Mulφ∗ (M) (V1 , ..., Vn , X, W1 , ..., Wm ; Y).

So the latter adjunction restricts to an adjunction V BEnrW ⇄ Ξ. To see that the adjunction is
an equivalence we prove that the right adjoint is conservative and the unit of the adjunction is an
equivalence. A V′ , W-enriched functor F ∶ M⊛ → N⊛ that is a morphism of Ξ is an equivalence if it
induces an equivalence on underlying ∞-categories and for every X, Y ∈ M and W1 , ..., Wm ∈ W for
m ≥ 0 the induced morphism LMulMorM (X, W1 , ..., Wm ; Y) → LMulMorN (F(X), W1 , ..., Wm ; F(Y))
in V′ is an equivalence. The latter morphism belongs to the essential image of φ and so is an equivalence
if for every V ∈ V the induced map
W(φ(V), LMulMorM (X, W1 , ..., Wm ; Y)) → W(φ(V), LMulMorN (F(X), W1 , ..., Wm ; F(Y)))
is an equivalence. The latter map identifies with the map
MulM (φ(V), X, W1 , ..., Wm ; Y) → MulN (φ(V), F(X), W1 , ..., Wm ; F(Y)),
which identifies with the map Mulφ∗ (M) (V, X, W1 , ..., Wm ; Y) → Mulφ∗ (N) (V, F(X), W1 , ..., Wm ; F(Y)).
This proves conservativity of φ∗ .
It remains to see that the unit is an equivalence. For every bi-enriched ∞-category M⊛ → V⊗ × W⊗
the unit α ∶ M⊛ → φ∗ (φ! (M))⊛ induces an essentially surjective functor M → φ! (M) on underlying
∞-categories by Proposition 4.61. So it remains to see that the unit induces an equivalence on left
morphism objects. By Proposition 4.61 for every W1 , ..., Wm ∈ W for m ≥ 0 and X, Y ∈ M the induced
morphism
φ(LMulMorM (X, W1 , ..., Wm ; Y)) → LMulMorφ! (M) (α(X), W1 , ..., Wm ; α(Y))
in V′ is an equivalence. For every V the unit α induces a map
V(V, LMulMorM (X, W1 , ..., Wm ; Y)) → V(V, LMulMorφ∗ (φ! (M)) (α(X), W1 , ..., Wm ; α(Y))) ≃
V′ (φ(V), LMulMorφ! (M) (α(X), W1 , ..., Wm ; α(Y))) ≃ V′ (φ(V), φ(LMulMorM (X, W1 , ..., Wm ; Y))),
which is the canonical map induced by φ and so is an equivalence by assumption.

4.6. Transforming bi-enrichment into left enrichment. It is well-known that for any monoidal
∞-category O⊗ → Ass and associative algebras A, B in O there is an equivalence between the ∞-
category of A, B-bimodules in O and the ∞-category of left A ⊗ Brev -modules in O (see for example
[14, Proposition 3.6.7., Corollary 3.6.8.] or [16, Theorem 4.3.2.7.]). Specializing to O = Cat∞ one
obtains for any small monoidal ∞-categories V⊗ → Ass, W⊗ → Ass an equivalence
(14) V BModW ≃ V×Wrev LMod∅
between the ∞-category of small ∞-categories bitensored over V, W and the ∞-category of small ∞-
categories left tensored over V × Wrev (see also Lemma 4.69). In this section we extend equivalence
(4.75) to enriched and pseudo-enriched ∞-categories (Theorem 4.74, Corollary 4.80, Corollary 4.88).
Moreover we prove that under this equivalence the respective ∞-categories of enriched presheaves
correspond (Proposition 4.94).
We start with transforming bi-pseudo-enrichment into left pseudo-enrichment (Theorem 4.74). For
that we first transform bitensored ∞-categories into left tensored ∞-categories. We give a proof of
that in our setting for the reader’s convenience (Lemma 4.69).
Notation 4.67. Let τ ∶= (id, (−)op ) ∶ Ass Ð
→ Ass × Ass and κ the functor Op∞ × Op∞ → Op∞ , (V⊗ →
Ass, W → Ass) ↦ V ×Ass (W ) . Taking pullback along τ defines a commutative square
⊗ ⊗ ⊗ rev
ωBEnr / ωLEnr
Op∞ × Op∞
κ

/ Op∞
corresponding to a functor
θ ∶ ωBEnr → κ∗ (ωLEnr)
over Op∞ × Op∞ .
Remark 4.68. The functor τ sends the map {0} ⊂ [n] in ∆ to the map ({0} ⊂ [n], {n} ⊂ [n]) in ∆×∆.
So for every weakly bi-enriched ∞-category M⊛ → V⊗ ×W⊗ and V1 , ..., Vn ∈ V, W1 , ..., Wn ∈ W, X, Y ∈ M
there is a canonical equivalence
Mulθ(M) ((V1 , W1 ), ..., (Vn , Wn ), X; Y) ≃ MulM (V1 , ..., Vn , X, W1 , ..., Wn ; Y).
So if M → V⊗ × W⊗ is a bitensored ∞-category, θ(M)⊛ → V⊗ ×Ass (W⊗ )rev provides a left V × Wrev -
⊛
action on M that sends (V, W) ∈ V × W, X ∈ M to V ⊗ X ⊗ W.

Lemma 4.69. The functor θ ∶ ωBEnr → κ∗ (ωLEnr) restricts to an equivalence
BMod ≃ κ∗ (LMod).
Proof. By Remark 4.68 the functor θ ∶ ωBEnr → κ∗ (ωLEnr) restricts to a functor BMod ≃ κ∗ (LMod),
which is a map of cartesian fibrations over Mon × Mon by the description of cartesian morphisms.
Consequently, it is enough to see that for any monoidal ∞-categories V⊗ → Ass, W⊗ → Ass the
induced functor θ ∶ ωBEnrV,W → ωLEnrV×Wrev restricts to an equivalence BModV,W ≃ LModV×Wrev .
By Proposition 2.51 the forgetful functors LModV×Wrev → Cat∞ , BModV,W → Cat∞ admit left adjoints
sending an ∞-category K to V⊛ ×Ass (W⊛ )rev × K → Ass, V⊛ × K × W⊛ → Ass × Ass, respectively, and
preserves small colimits by Lemma 2.52. Thus these functors are monadic by [16, Theorem 4.7.3.5].
So by [16, Corollary 4.7.3.16.] the claim follows from the fact that θ preserves the left adjoints, i.e.
the canonical functor V⊛ ×Ass (W⊛ )rev × K → θ(V⊛ × K × W⊛ ) is an equivalence.
Notation 4.70. Let M⊛ → V⊗ × W⊗ be a bi-pseudo-enriched ∞-category. Let
̃
BEnv(M)⊛
⊂ V⊗ ×P(V)⊗ PBEnvBPseu (M)⊛ ×P(W)⊗ W⊗
be the full subcategory with V, W-biaction generated by M, i.e. the full weakly bitensored subcategory
spanned by the objects V ⊗ X ⊗ W for V ∈ V, W ∈ W, X ∈ M.
Lemma 4.71. Let M⊛ → V⊗ × W⊗ be a bi-pseudo-enriched ∞-category. The V, W-linear embedding
̃
BEnv(M)⊛
⊂ V⊗ ×P(V)⊗ PBEnvBPseu (M)⊛ ×P(W)⊗ W⊗ induces a P(V), P(W)-linear equivalence
̃
P(BEnv(M))⊛
≃ PBEnvBPseu (M)⊛ .
Proof. By Lemma 4.46 for every V ∈ V, W ∈ W, X ∈ M the object V ⊗ X ⊗ W is ∅-compact in
PBEnvBPseu (M). So by [17, Corollary 5.1.6.11.] the result follows from the fact that PBEnvBPseu (M)
is generated under small colimits by the objects V ⊗ X ⊗ W for V ∈ V, W ∈ W, X ∈ M.
Proposition 4.72. Let M⊛ → V⊗ × W⊗ be a bi-pseudo-enriched ∞-category and N⊛ → V⊗ × W⊗ a
bitensored ∞-category. The lax V, W-linear embedding M⊛ ⊂ BEnv(M)
̃ ⊛
ψN ∶ LinFunV,W (BEnv(M),
̃ N) → LaxLinFunV,W (M, N).
Proof. By Proposition 2.104 the ∞-category N⊛ → V⊗ × W⊗ bitensored over V, W embeds into an
∞-category P(N)⊛ → P(V)⊗ × P(W)⊗ bitensored over P(V), P(W) compatible with small colimits,
and ψN is the pullback of ψP(N) . So it is enough to check that ψP(N) is an equivalence. Consider the
commutative triangle:
LinFunLP(V),P(W)(PBEnvBPseu (M), P(N))
❲❲❲❲❲
❲❲❲❲❲ρ
❲❲❲❲❲
❲❲❲❲❲
❲+
LaxLinFunV,W (M, P(N))
ψP(N) ❢❢❢❢❢❢❢
❢❢3
❢❢ ❢
❢❢❢❢
❢❢❢❢❢
̃
LinFunV,W (BEnv(M) ⊛
, P(N))
By Proposition 4.20 the functor ρ is an equivalence. By Lemma 4.71 the vertical functor is an
equivalence.

Corollary 4.73. Let M⊛ → V⊗ ×W⊗ be a weakly bitensored ∞-category and M̄⊛ → Env(V)⊗ ×Env(W)⊗
the unique extension to a bi-pseudo-enriched ∞-category. The embedding M̄⊛ ⊂ BEnv(M)⊛ induces
an Env(V), Env(W)-linear equivalence
̃ M̄)⊛ ≃ BEnv(M)⊛ .
BEnv(
Proof. For every bitensored ∞-category N⊛ → Env(V)⊗ × Env(W)⊗ by Proposition 4.72 the functor
̃ M̄), N) → LaxLinFunEnv(V),Env(W) (M̄, N) ≃ LaxLinFunV,W (M, N)
LinFunEnv(V),Env(W) (BEnv(
is an equivalence, where the latter equivalence is by Corollary 4.32.

Theorem 4.74. The functor θ ∶ BPEnr → κ∗ (LPEnr∅ ) is an equivalence.
Proof. The functor θ ∶ BPEnr → κ∗ (LPEnr∅ ) is a map of cocartesian fibrations over Op∞ × Op∞
using Proposition 4.61 (4). So it is enough to see that for any ∞-operads V⊗ → Ass, W⊗ → Ass
the induced functor θ ∶ V BPEnrW → V×Wrev LPEnr∅ on the fiber is an equivalence. We construct an
inverse γ of θ. By Proposition 4.72 and Corollary ?? (1) the inclusion V×Wrev LMod∅ ⊂ V×Wrev LPEnr∅
admits a left adjoint L. By Lemma 4.69 the functor θ ∶ V BPEnrW → V×Wrev LPEnr∅ restricts to
an equivalence V BModW ≃ V×Wrev LMod, whose inverse we call Λ. For every bi-pseudo-enriched ∞-
category M⊛ → V⊗ × (Wrev )⊗ let γ(M)⊛ ⊂ Λ(L(M))⊛ be the full weakly bi-enriched subcategory
spanned by M ⊂ L(M). Since Λ(L(M))⊛ → V⊗ × W⊗ is a bitensored ∞-category, γ(M)⊛ → V⊗ × W⊗

exhibits M as bi-pseudo-enriched in V, W. Thus the functor Λ ○ L ∶ V×Wrev LPEnr∅ Ð → V×Wrev LMod∅ Ð →
V BMod W induces a functor γ ∶ V×W rev LPEnr∅ → V BPEnr W . We will prove that γ is inverse to θ.
The tautological natural transformation γ ⊂ Λ ○ L of functors V×Wrev LPEnr∅ → V BPEnrW gives rise
to a natural transformation θ ○ γ ⊂ θ ○ Λ ○ L ≃ L of functors V×Wrev LPEnr∅ → V×Wrev LPEnr∅ , which
induces an equivalence α ∶ θ ○ γ ≃ id of functors V×Wrev LPEnr∅ → V×Wrev LPEnr∅ . By Proposition 4.72
the inclusion V BModW ⊂ V BPEnrW admits a left adjoint L′ . The composite L ○ θ → L ○ θ ○ L′ ≃ θ ○ L′
(induced by the unit id → L′ ) of natural transformations of functors V BPEnrW → V×Wrev LMod∅ gives
rise to a natural transformation Λ ○ L ○ θ → Λ ○ θ ○ L′ ≃ L′ of functors V BPEnrW → V BModW . Let β be
the composition γ ○ θ ⊂ Λ ○ L ○ θ → L′ of natural transformations of functors V BPEnrW → V BPEnrW .
There is a commutative triangle:
θ ○ γ ○●θ / θ ○ L′
θ○β
●● ③=
●●α○θ ③③③
●● ③
●● ③③
# ③③
θ.
Since α is an equivalence and the unit id → L′ is component-wise an embedding, θ ○ β and so β
are component-wise embeddings. Hence β ∶ γ ○ θ → L′ induces an equivalence γ ○ θ ≃ id of functors
V BPEnrW → V BPEnrW .

Corollary 4.75. Let V⊗ → Ass, W⊗ → Ass be small monoidal ∞-categories. There is a canonical
equivalence
V PBEnrW ≃ V×Wrev PLEnr∅ .
Corollary 4.76. Let V⊗ → Ass, W⊗ → Ass be small ∞-operads. There is a canonical equivalence
V ωBEnrW ≃ PEnv(V)⊗PEnv(Wrev ) LEnr∅ .
Proof. By Theorem 4.74 there is a canonical equivalence
Env(V) BPEnrEnv(W) ≃ Env(V)×Env(W)rev LPEnr∅ .
By Corollary 4.32 there are canonical equivalences
V ωBEnrW ≃ Env(V) BPEnrEnv(W) , P(Env(V)×Env(W)rev ) LEnr∅ ≃ Env(V)×Env(W)rev LPEnr∅ .
Corollary 4.77. For every bi-pseudo-enriched ∞-category M⊛ → V⊗ × W⊗ and weakly bi-enriched

∞-category N⊛ → U⊗ × Q⊗ the following induced commutative square is a pullback square:
(15) EnrFun(M, N) / EnrFun(θ(M), θ(N))
AlgV (U) × AlgW (Q) V×Wrev (U × Q

rev
/ Alg ).
Proof. The commutative square (15) is the pullback of the commutative square (15), where we re-
place N⊛ → U⊗ × Q⊗ by the enveloping bitensored ∞-category BEnv(N)⊛ → Env(U)⊗ × Env(Q)⊗ .
Consequently, we can assume that N⊛ → U⊗ × Q⊗ is a bi-pseudo-enriched ∞-category. For any small
∞-category K applying the functor Cat∞ (K, −) ∶ Cat∞ → S to square (15) gives the induced commu-
tative square
ωBEnr(M, NK ) / ωLEnr(θ(M), θ(NK ))

Op∞ (V, UK ) × Op∞ (W, QK ) / Op (V × Wrev , UK × (QK )rev ).
∞
So the result follows from Theorem 4.74 and that (M⊛ )K → (V⊗ )K × (W⊗ )K is a bi-pseudo-enriched
∞-category.

Remark 4.78. The functor θ ∶ ωBEnr → κ∗ (ωLEnr) is not an equivalence since it is not conservative.
There is a canonical commutative square:
θ
Env(V) BPEnrEnv(W) Env(V)×Env(W)rev LPEnr∅
/
≃
≃
θ

V ωBEnrW
/ V×Wrev ωLEnr ≃ Env(V×Wrev ) LPEnr∅ ,
which identifies the functor θ ∶ V ωBEnrW → V×Wrev ωLEnr with the right vertical functor
Env(V)×Env(W)rev LPEnr∅ → Env(V×Wrev ) LPEnr∅
taking pullback along the monoidal functor Env(V × Wrev )⊗ → Env(V)⊗ ×Ass Env(Wrev )⊗ .
We apply Theorem 4.74 to transform bi-enrichment into left enrichment (Corollary 4.80).
Notation 4.79. Let V⊗ → Ass, W⊗ → Ass be small ∞-operads.
(1) Let Fun(Env(V)op , W)⊗ ⊂ P(Env(V) × Env(W))⊗ be the full suboperad whose colors are the
objects of Fun(Env(V)op , W) ⊂ Fun(Env(V)op , PEnv(W)) ≃ P(Env(V) × Env(W)).
(2) Let ⟨V, W⟩⊗ ⊂ P(Env(V)×Env(W))⊗ be the full suboperad whose colors are the presheaves on
Env(V) × Env(W) representable in both variables such that for the corresponding adjunction
Env(V) ⇄ Env(W)op the left adjoint lands in Wop and the right adjoint lands in V.
Corollary 4.80. Let V⊗ → Ass, W⊗ → Ass be small ∞-operads. There are canonical equivalences
preserving the underlying ∞-category and morphism objects:
(1)
V LEnrW ≃ Fun(Env(Wrev )op ,V) LEnr,
(2)
V REnrW ≃ Fun(Env(V)op ,Wrev ) LEnr,
(3)
V BEnrW ≃ ⟨V,Wrev ⟩ LEnr.
Proof. (1): Under the equivalence V ωBEnrW ≃ P(Env(V)×Env(W)rev ) LEnr of Corollary 4.76 the full
subcategory V LEnrW corresponds to the full subcategory Ξ ⊂ P(Env(V)×Env(W)rev ) LEnr spanned by
the left enriched ∞-categories M⊛ → P(Env(V) × Env(W)rev )⊗ such that for every X, Y ∈ M the left
morphism object LMorM (X, Y) ∈ P(Env(V) × Env(W)) belongs to
Fun(Env(W)op , V) ⊂ Fun(Env(W)op , PEnv(V)) ≃ P(Env(V) × Env(W)).
The induced functor P(Env(V)×Env(W)rev ) ωLEnr → Fun(Env(Wrev )op ,V) ωLEnr restricts to a functor
Ξ → Fun(Env(Wrev )op ,V) LEnr,
which is an equivalence by Corollary 4.66. (2) and (3) are similar to (1).

Corollary 4.80 gives the following theorem as a corollary:
Theorem 4.81. Let V⊗ → Ass, W⊗ → Ass be presentably monoidal ∞-categories. There is a canonical
equivalence preserving the underlying ∞-category and morphism objects:
V BEnrW ≃ V⊗Wrev LEnr.
Next we apply Theorem 4.74 to tranform bi-modules into left modules (Corollaries 4.84, ??).
Notation 4.82. Let M⊛ → V⊗ be a weakly left enriched ∞-category, O⊛ → W⊗ a weakly right
enriched ∞-category, N⊛ → V⊗ × W⊗ a weakly bi-enriched ∞-category and A, B associative algebras
in V, W, respectively. Let
A M → Ass, OB → Ass, A NB → Ass × Ass
⊛ ⊛ ⊛
be the pullbacks of M⊛ → V⊗ along A ∶ Ass → V⊗ , of O⊛ → W⊗ along B ∶ Ass → W⊗ and of

N⊛ → V⊗ × W⊗ along A × B ∶ Ass × Ass → V⊗ × W⊗ , respectively. Let
A LMod(M) ∶= EnrFun[0] ([0], A M), LMod(O)B ∶= EnrFun[0] ([0], OB ),
A BModB (N) ∶= EnrFun[0],[0] ([0], A NB ).
We apply Notation 4.82 to V⊛ → V⊗ × V⊗ and ∅⊗ ×V⊗ V⊛ → V⊗ for any ∞-operad V⊗ → Ass.
Remark 4.83. Let N⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category and A, B associative algebras
in V, W, respectively. There is a canonical equivalence
A LMod(RModB (N)) = EnrFun[0] ([0], A EnrFun[0] ([0], NB )) ≃ EnrFun[0] ([0], EnrFun[0] ([0], A NB )) ≃
EnrFun[0],[0] ([0], A NB ) = A BModB (N).
Corollary 4.84. Let N⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category, A an associative algebra in
V and B an associative algebra in W. The following canonical functor is an equivalence:
A BModB (N) → (A,Brev ) LMod(θ(N)).
Next we apply Theorem 4.74 to show that (κ, κ)-bi-pseudo-enriched ∞-categories can be trans-
formed into left κ-pseudo-enriched ∞-categories for any regular cardinal κ (Corollary 4.88).
Notation 4.85. Let n, m ≥ 0 and for every 1 ≤ i ≤ n, 1 ≤ j ≤ m let V⊗ i → Ass, Wj → Ass be small
⊗
monoidal ∞-categories compatible with κi -small colimits, τj -small colimits, respectively, for small
regular cardinals κi , τj . Let
κ1 ,...,κn τ1 ,...,τm
V1 ×...×Vn BEnrW1 ×...×Wm ⊂ V1 ×...×Vn BPEnrW1 ×...×Wm
be the full subcategory spanned by the bi-pseudo-enriched ∞-categories whose morphism objects are
presheaves on V1 ×...×Vn ×W1 ×...×Wm preserving κi -small colimits in the i-th component for 1 ≤ i ≤ n
and preserving τj -small colimits in the n + j-th component for 1 ≤ j ≤ m.
i →
Lemma 4.86. Let κ, τ be small regular cardinals, n, m ≥ 0 and for every 1 ≤ i ≤ n, 1 ≤ j ≤ m let V⊗
Ass, W⊗j → Ass be small monoidal ∞-categories compatible with κ-small colimits, τ -small colimits,
respectively. The functor
V1 ⊗κ ...⊗κ Vn BEnrW1 ⊗τ ...⊗τ Wm → V1 ×...×Vn BPEnrW1 ×...×Wm
taking pullback along the universal monoidal functors preserving κ-small colimits, τ -small colimits
component-wise, respectively, induces an equivalence
→ κ,...,κ
κ τ τ,...,τ
V1 ⊗κ ...⊗κ Vn BEnrW1 ⊗τ ...⊗τ Wm V1 ×...×Vn BEnrW 1 ×...×Wm
.
Proof. The universal monoidal functor α ∶ V⊗

1 ×Ass ...×Ass Vn → (V1 ⊗κ ...⊗κ Vn ) induces an adjunction
⊗ ⊗
α! ∶ (P(V1 ) ⊗ ... ⊗ P(Vn ))⊗ ≃ P(V1 × ... × Vn )⊗ → P(V1 ⊗κ ... ⊗κ Vn )⊗ ∶ α∗

relative to Ass. By universal property of α the latter adjunction restricts to a monoidal equivalence
α! ∶ (Indκ (V1 ) ⊗ ... ⊗ Indκ (Vn ))⊗ ≃ Indκ (V1 ⊗κ ... ⊗κ Vn )⊗ ∶ α∗ ,
where (Indκ (V1 ) ⊗ ... ⊗ Indκ (Vn ))⊗ ⊂ P(V1 × ... × Vn )⊗ is the full suboperad spanned by the presheaves
preserving component-wise κ-small colimits, which is a localization relative to Ass and so a presentably
monoidal ∞-category. By Corollary 4.34 the canonical functor
Indκ (V1 ⊗κ ...⊗κ Vn ) BEnrIndτ (W1 ⊗τ ...⊗τ Wm ) → κV1 ⊗κ ...⊗κ Vn BEnrτW1 ⊗τ ...⊗τ Wm
is an equivalence. So it is enough to see that the composition
→ κV1 ⊗κ ...⊗κ Vn BEnrτW1 ⊗τ ...⊗τ Wm → κ,...,κ
τ,...,τ
Indκ (V1 ⊗κ ...⊗κ Vn ) BEnrIndτ (W1 ⊗τ ...⊗τ Wm ) V1 ×...×Vn BEnrW 1 ×...×Wm
is an equivalence. The latter composition factors as
Indκ (V1 ⊗κ ...⊗κ Vn ) BEnrIndτ (W1 ⊗τ ...⊗τ Wm ) ≃ Indκ (V1 )⊗...⊗Indκ (Vn ) BEnrIndτ (W1 )⊗...⊗Indτ (Wm ) →
κ,...,κ τ,...,τ
V1 ×...×Vn BEnrW1 ×...×Wm .
Let Θ ⊂ P(V1 ×...×Vn ) BEnrP(W1 ×...×Wm) be the full subcategory of bi-enriched ∞-categories whose mor-
phism objects belong to Indκ (V1 )⊗...⊗Indκ (Vn )⊗Indτ (W1 )⊗...⊗Indτ (Wm ) ⊂ P(V1 ×...×Vn ×W1 ×...×
Wm ). The functor Θ ⊂ P(V1 ×...×Vn ) BEnrP(W1 ×...×Wm) → Indκ (V1 )⊗...⊗Indκ (Vn ) BPEnrIndτ (W1 )⊗...⊗Indτ (Wm )
induces an equivalence Θ → Indκ (V1 )⊗...⊗Indκ (Vn ) BEnrIndτ (W1 )⊗...⊗Indτ (Wm) by Proposition 4.66. So
it is enough to see that the functor P(V1 ×...×Vn ) BEnrP(W1 ×...×Wm) → V1 ×...×Vn BPEnrW1 ×...×Wm is an
τ,...,τ
equivalence that restricts to an equivalence Θ ≃ κ,...,κ
V1 ×...×Vn BEnrW1 ×...×Wm . This holds by Corollary 4.32.

Corollary 4.87. Let κ be a small regular cardinal and V⊗ → Ass, W⊗ → Ass small monoidal ∞-
categories compatible with κ-small colimits. The following induced functor is an equivalence:
κ
V⊗κ W LEnr∅ → κ,κ
V×W LEnr∅ .
Theorem 4.74 gives the following corollary:

Corollary 4.88. Let κ be a small regular cardinal and V⊗ → Ass, W⊗ → Ass small monoidal ∞-
categories compatible with κ-small colimits. The equivalence
V BPEnrW ≃ V×Wrev LPEnr∅

of Theorem 4.74 taking pullback along the projection V⊗ ×Ass (W⊗ )rev → V⊗ × (W⊗ )rev ≃ V⊗ × W⊗
θ ∶ κV BEnrκW → κ,κ κ
V×Wrev LEnr∅ ≃ V⊗κ Wrev LEnr∅ .
Corollary 4.89. Let V⊗ → Ass, W⊗ → Ass be presentably monoidal ∞-categories. There is a canonical
equivalence
V BEnrW ≃ V⊗Wrev LEnr∅ .
Next we prove that the ∞-category of enriched presheaves is invariant under transforming bi-
enrichment into left enrichment (Corollary 4.93). To state the following results we need the next
remark:
Remark 4.90. Let κ < κ′ be small regular cardinals, U⊗ → Ass a small ∞-operad, V⊗ → Ass, W⊗ →
Ass small monoidal ∞-categories compatible with κ-small colimits, V′⊗ → Ass, W′⊗ → Ass small
monoidal ∞-categories compatible with κ′ -small colimits and α ∶ V⊗ → V′⊗ , β ∶ W⊗ → W′⊗ monoidal
functors preserving κ-small colimits. Let γ ∶ (V ⊗κ W)⊗ → (V′ ⊗κ W′ )⊗ → (V′ ⊗κ′ W′ )⊗ be the induced
monoidal functor preserving κ-small colimits. There is a commutative square:
V′ ⊗κ′ W′ LEnrU V′ ×W′ LEnrU

/
γ (α×β)∗

V⊗κ W LEnrU
/ V×W LEnrU .
This follows from naturality and the fact that the monoidal functor V′⊗ ×Ass W′⊗ → (V′ ⊗κ′ W′ )⊗
factors as V′⊗ ×Ass W′⊗ → (V′ ⊗κ W′ )⊗ → (V′ ⊗κ′ W′ )⊗ . Hence there is also a commutative square:
κ′ κ′ κ′
V′ BEnrW′ V′ ⊗W′rev LEnr∅
/
(α,β)∗ γ∗

κ κ / κ rev LEnr∅ .
V BEnrW V⊗W
Proposition 4.91. Let κ be a small regular cardinal and M⊛ → V⊗ × W⊗ a small κ, κ-bi-enriched

∞-category. The V ⊗κ Wrev -enriched embedding
θ(M)⊛ → θ(PBEnv(M)BEnrκ,κ )⊛
induces a left Indκ (V) ⊗ Indκ (W)rev ≃ Indκ (V ⊗κ Wrev )-linear equivalence
LEnrκ → θ(PBEnv(M)BEnrκ,κ ) .
PLEnv(θ(M))⊛ ⊛
Proof. The left Indκ (V ⊗κ Wrev )-linear functor γ ∶ PLEnv(θ(M))⊛ LEnrκ → θ(PBEnv(M)BEnrκ,κ ) is
⊛
an equivalence if the pullback of γ to (V ⊗κ W ) is an equivalence because source and target of γ

rev ⊗
are left V × Wrev -enriched ∞-categories. By Corollary 4.88 the functor θ ∶ κV BEnrκW → κ,κ
V×Wrev LEnr∅ ≃
κ
V⊗W rev LEnr∅
taking pullback along the projection V⊗
× Ass (W ⊗ rev
) → V ⊗
× (W ⊗ rev
) ≃ V⊗ × W⊗ is
an equivalence. The pullback of γ to (V ⊗κ Wrev )⊗ is an embedding γ ′ ∶ PLEnv ̃ LEnrκ (θ(M))⊛ →
θ(PBEnv(M)BEnrκ,κ ) . Let N → V ×Ass (W )
̃ ⊛ ⊛ ⊗ ⊗ rev
be a left tensored ∞-category that admits small
conical colimits. Consider the commutative square:
̃
LinFunLV,W (PBEnv(M) −1
BEnrκ,κ , θ (N))
≃ / EnrFunV,W (M, θ−1 (N))
≃ ≃

̃
LinFunLV⊗κ Wrev (θ(PBEnv(M)BEnrκ,κ ), N)
/ EnrFunV⊗ Wrev (θ(M), N).
κ
Corollary 4.92. Let κ be a small regular cardinal and M⊛ → V⊗ × W⊗ a small κ, κ-bi-enriched

∞-category. The left V ⊗κ Wrev -enriched embedding θ(M)⊛ → θ(PBEnvκ (M)BEnr )⊛ induces a left
V ⊗κ Wrev -linear equivalence
LEnr → θ(PBEnvκ (M)BEnr ) .

PLEnvκ (θ(M))⊛ ⊛
Corollary 4.93. Let M⊛ → V⊗ × W⊗ be a small bi-enriched ∞-category. The left V ⊗ Wrev -enriched
embedding θ(M)⊛ → θ(PBEnv(M)BEnr )⊛ induces a left V ⊗ Wrev -linear equivalence
LEnr → θ(PBEnv(M)BEnr ) .
PLEnv(θ(M))⊛ ⊛
Corollary 4.94. Let κ be a small regular cardinal, M⊛ → V⊗ × W⊗ a totally small weakly bi-enriched
∞-category and M̄⊛ → (PEnv(V)κ )⊗ × (PEnv(W)κ )⊗ the unique extension to a κ, κ-bi-enriched ∞-
category. The left PEnv(V)κ ⊗κ (PEnv(W)κ )rev -enriched embedding
θ(M̄)⊛ → θ(PBEnv(M))⊛
induces a left Indκ (PEnv(V)κ ⊗κ (PEnv(W)κ )rev ) ≃ PEnv(V) ⊗ PEnv(W)rev -linear equivalence
LEnrκ → θ(PBEnv(M)) .
PLEnv(θ(M̄))⊛ ⊛
5. Enriched ∞-categories of enriched functors

In this section we introduce and study enriched ∞-categories of enriched functors.
Notation 5.1. Let V⊗ → Ass, W⊗ → Ass be small ∞-operads. The functor
Cat∞/V⊗ × Cat∞/W⊗ → Cat∞/V⊗ ×W⊗ , (M⊛ → V⊗ , N⊛ → W⊗ ) ↦ M⊛ × N⊛ → V⊗ × W⊗
restricts to a functor
(16) α ∶ V ωBEnr∅ × ∅ ωBEnrW → V ωBEnrW .
Remark 5.2. Let V → Ass, W⊗ → Ass be small monoidal ∞-categories. The functor (16) restricts
⊗
to a functor
(17) V PLEnr∅ × ∅ PREnrW → V PBEnrW .
We have the following proposition, which follows immediately from Proposition 5.11:
Proposition 5.3. For every small ∞-operads V⊗ → Ass, W⊗ → Ass the functor (16) admits compo-
nentwise right adjoints.
We will prove later that also the functor (17) admits componentwise right adjoints (Corollary 5.53).
To prove Proposition 5.11 we use the following lemma:
Lemma 5.4. Let C → T, T → S be functors such that the composition C → T → S is a cocartesian
fibration. The functor (−) ×S C ∶ Cat∞/S Ð → Cat∞/C → Cat∞/T admits a right adjoint that we denote by
FunST (C, −). If S, T are contractible, we drop S, T from the notation.
Proof. It is enough to see that the functor C ×S (−) ∶ Cat∞/S → Cat∞/C admits a right adjoint. This
follows from [16, Example B.3.11.] and [16, Corollary B.3.15.].

Remark 5.5. For every functors D → T and S′ → S there is a canonical equivalence S′ ×S FunST (C, D) ≃
FunSS′ ×S T (S′ ×S C, S′ ×S D) specifying the fibers of the functor FunST (C, D) → S.
′
Remark 5.6. Let φ ∶ T → S be a functor, E ⊂ Fun([1], S) a full subcategory and C → T a cartesian

fibration relative to the collection of φ-cocartesian lifts of morphisms of S that belong to E and D → T
a cocartesian fibration relative to the collection of φ-cocartesian lifts of morphisms of S that belong
to E. By [16, Proposition B.4.1.] the functor FunST (C, D) → S is a cocartesian fibration relative to E.
Lemma 5.7. Let C → T, T → S be functors such that the composition C → T → S is a cocartesian
fibration. Let D → E be a fully faithful functor over T. Then FunST (C, D) → FunST (C, E) is fully faithful.
Proof. Note that the monomorphisms in Cat∞/T are detected by the forgetful functor to Cat∞ and
so are the functors over T whose image in Cat∞ is an inclusion, i.e. induces embeddings on mapping
spaces and the maximal subspace. Because the functor FunST (C, −) ∶ Cat∞/T → Cat∞/S is a right
adjoint, it preserves monomorphisms. So it remains to see that the functor FunST (C, D) → FunST (C, E)
is full, i.e. essentially surjective on mapping spaces. Let φ ∶ [1] → FunST (C, E) be a functor over S
whose restrictions to {0} ⊂ [1], {1} ⊂ [1] belong to FunST (C, D). Then φ corresponds to a functor
ψ ∶ [1] ×S C → [1] ×S E over [1] whose fibers over 0, 1 factor through D. But then ψ factors through D
so that φ factors through FunST (C, D).

The following lemma is [10, Lemma 3.73.] whose proof we include for the reader’s convenience:
Lemma 5.8. Let M⊛ → V⊗ , O⊛ → W⊗ , N⊛ → V⊗ × W⊗ be weakly left, weakly right, weakly bi-enriched
∞-categories.
(1) The functor α ∶ FunW
V⊗ ×W⊗ (M × W , N ) → W is a weakly right enriched ∞-category.
⊗
⊛ ⊗ ⊛ ⊗
(2) The functor β ∶ FunV

V⊗ ×W⊗ (V × O , N ) → V is a weakly left enriched ∞-category.
⊗
⊗ ⊛ ⊛ ⊗
(3) If N⊛ → V⊗ × W⊗ exhibits N as right tensored over W, the functor α is a right tensored

∞-category and for every X ∈ M the following canonical functor is right W-linear:
V⊗ ×W⊗ (M × W , N ) → N .
⊗
⊛ ⊗ ⊛ ⊛
FunW
(4) If N⊛ → V⊗ ×W⊗ exhibits N as left tensored over V, the functor β is a left tensored ∞-category
and for every X ∈ O the following canonical functor is left V-linear:
V⊗ ×W⊗ (V × O , N ) → N .
⊗
⊗ ⊛ ⊛ ⊛
FunV
Proof. We prove (1) and (3). (2) and (4) are dual. By Remark 5.6 the functor α and for every
V⊗ ×W⊗ (M × W , N ) → N are maps of cocartesian fibrations relative to the
⊗
X ∈ M the functor FunW ⊛ ⊗ ⊛ ⊛
collection of inert morphisms of Ass preserving the maximum. Moreover these functors are maps of
cocartesian fibrations over Ass if N⊛ → V⊗ × W⊗ exhibits N as right tensored over W. For the general
case every cocartesian lift W → W′ in W⊗ of the map {n} ⊂ [n] in Assop induces the functor
V⊗ ×W⊗ (M × W , N )W ≃ FunV⊗ (M , NW ) → FunV⊗ (M , NW′ ) ≃ FunV⊗ ×W⊗ (M × W , N )W′

⊗ ⊗
⊛ ⊗ ⊛ ⊛ ⊛ ⊛ ⊛ ⊛ ⊗ ⊛
FunW W
induced by the equivalence NW ⊛

→ NW⊛
′ . This proves (2). To see (1) it remains to verify condition (2) of
V⊗ ×W⊗ (M × W , REnv(N) ) → Env(W) is a right tensored

⊗
Definition 2.22. By (2) the functor FunW ⊛ ⊗ ⊛ ⊗
∞-category. So the pullback W ×Env(W)⊗ FunV⊗ ×W⊗ (M × W , REnv(N) ) → W⊗ is a weakly right

⊗ W⊗ ⊛ ⊗ ⊛
enriched ∞-category. By Lemma 5.7 the embedding N⊛ → REnv(N)⊛ gives rise to an embedding
V⊗ ×W⊗ (M × W , N ) ↪ W ×Env(W)⊗ FunV⊗ ×W⊗ (M × W , REnv(N) )

⊗ ⊗
⊛ ⊗ ⊛ ⊗ ⊛ ⊗ ⊛
FunW W
over W⊗ that is a map of cocartesian fibrations relative to the collection of inert morphisms of Ass
V⊗ ×W⊗ (M × W , N ) → W satisfies condition (2)
⊗
preserving the maximum. This implies that FunW ⊛ ⊗ ⊛ ⊗
of Definition 2.22 and so is a right enriched ∞-category.

The next notation is [10, Notation 3.74.]:
Notation 5.9. Let M⊛ → V⊗ , O⊛ → W⊗ , N⊛ → V⊗ × W⊗ be weakly left, weakly right and weakly
bi-enriched ∞-categories, respectively.
(1) Let
V⊗ ×W⊗ (M × W , N )
⊗
EnrFunV,∅ (M, N)⊛ ⊂ FunW ⊛ ⊗ ⊛
be the full weakly right W-enriched subcategory spanned by
V⊗ ×W⊗ (M × W , N )[0] .
⊗
EnrFunV,∅ (M, N) ⊂ FunV⊗ (M⊛ , N[0]
⊛
) ≃ FunW ⊛ ⊗ ⊛
(2) Let
V⊗ ×W⊗ (V × O , N )
⊗
EnrFun∅,W (O, N)⊛ ⊂ FunV ⊗ ⊛ ⊛
be the full weakly left V-enriched subcategory spanned by
V⊗ ×W⊗ (M × W , N )[0] .
⊗
EnrFunV,∅ (M, N) ⊂ FunV⊗ (M⊛ , N[0]
⊛
) ≃ FunW ⊛ ⊗ ⊛
Remark 5.10. There is a canonical left V-enriched equivalence:

EnrFun∅,W (O, N)⊛ ∶= (EnrFunW,∅ (Orev , Nrev )rev )⊛ .
Remark 5.6 implies the following proposition, which is [10, Proposition 3.78.]:
Proposition 5.11. Let M⊛ → V⊗ , O⊛ → W⊗ , N⊛ → V⊗ × W⊗ be weakly left, weakly right and weakly
bi-enriched ∞-categories, respectively.
(1) The canonical equivalence
V⊗ ×W⊗ (M × W , N )) ≃ FunV⊗ ×W⊗ (M × O , N )

⊗
FunW⊗ (O⊛ , FunW ⊛ ⊗ ⊛ ⊛ ⊗ ⊛
EnrFun∅,W (O, EnrFunV,∅ (M, N)) ≃ EnrFunV,W (M × O, N).
(2) The canonical equivalence
V⊗ ×W⊗ (V × O , N )) ≃ FunV⊗ ×W⊗ (M × O , N )

⊗
FunV⊗ (M⊛ , FunV ⊗ ⊛ ⊛ ⊛ ⊗ ⊛
EnrFunV,∅ (M, EnrFun∅,W (O, N)) ≃ EnrFunV,W (M × O, N).
Remark 5.5 implies the following remark:
Remark 5.12. Let M⊛ → V⊗ , N⊛ → V⊗ × W⊗ be weakly left, weakly bi-enriched ∞-categories,
respectively. For every map W′⊗ → W⊗ of ∞-operads there is a canonical right W′ -enriched equivalence
W′⊗ ×W⊗ EnrFunV,∅ (M, N)⊛ ≃ EnrFunV,∅ (M, W′ ×W N)⊛ .
Lemma 5.7 implies the following remark:
Remark 5.13. Let M⊛ → V⊗ , N⊛ → V⊗ × W⊗ , N′⊛ → V⊗ × W⊗ be weakly left, weakly bi-enriched ∞-
categories, respectively. For every V, W-enriched embedding N⊛ → N′⊛ the induced right W-enriched
functor EnrFunV,∅ (M, N)⊛ → EnrFunV,∅ (M, N′ )⊛ is an embedding.
Lemma 5.8, Remark 5.6 and Lemma 2.52 imply the following corollary:
Corollary 5.14. Let K ⊂ Cat∞ be a full subcategory and M⊛ → V⊗ , N⊛ → V⊗ ×W⊗ weakly left enriched
and weakly bi-enriched ∞-categories. If N⊛ → V⊗ × W⊗ is a right tensored ∞-category compatible
with K-indexed colimits, EnrFunV,∅ (M, N)⊛ → W⊗ is a right tensored ∞-category compatible with
K-indexed colimits and the right W-enriched functor EnrFunV,∅ (M, N)⊛ → (NM )⊛ is right W-linear
and preserves K-indexed colimits.
The latter corollary admits the following refinement:
Proposition 5.15. Let K ⊂ Cat∞ be a full subcategory and M⊛ → V⊗ , N⊛ → V⊗ × W⊗ weakly left
enriched and weakly bi-enriched ∞-categories. If N⊛ → V⊗ ×W⊗ admits right tensors and forming right
tensors preserves K-indexed colimits, EnrFunV,∅ (M, N)⊛ → W⊗ admits right tensors and K-indexed
conical colimits and the right W-enriched functor EnrFunV,∅ (M, N)⊛ → (NM )⊛ is right W-linear and
preserves K-indexed colimits.
Proof. By Lemma 2.52 it is enough to prove the case that K is empty. By Corollary 4.38 (2) (for
κ = ∅) the weakly bi-enriched ∞-category N⊛ → V⊗ × W⊗ that admits right tensors is the pullback
of a unique right tensored ∞-category N′⊛ → V⊗ × Env(W)⊗ . By Proposition 5.8 the weakly right
enriched ∞-category EnrFunV,∅ (M, N′ )⊛ → Env(W)⊗ is a right tensored ∞-category and the right
Env(W)-enriched functor EnrFunV,∅ (M, N′ )⊛ → (N′M )⊛ is right linear. So using Remark 5.12 the
pullback EnrFunV,∅ (M, N)⊛ ≃ EnrFunV,∅ (M, N′ )⊛ ×Env(W)⊗ W⊗ → W⊗ is a weakly right enriched
∞-category that admits right tensors and the following right W-enriched functor is right W-linear:
EnrFunV,∅ (M, N)⊛ ≃ EnrFunV,∅ (M, N′ )⊛ ×Env(W)⊗ W⊗ → Fun(M, N)⊛ ≃ Fun(M, N′ )⊛ ×Env(W)⊗ W⊗ .

5.1. Enriched Kan-extensions. In the following we use Proposition 2.62 to construct a left Kan-
extension for ∞-categories enriched in any presentably monoidal ∞-category (Proposition 5.16, Propo-
sition 5.20). Passing to opposite enriched ∞-categories we obtain a formally dual theory of right
Kan-extensions for ∞-categories enriched in any presentably monoidal ∞-category.
Proposition 5.16. (1) Let V⊗ → Ass, W⊗ → Ass be small ∞-operads and ψ ∶ M⊛ → N⊛ a V, W-
enriched functor from a totally small to a locally small weakly bi-enriched ∞-category. Let
D⊛ → V⊗ ×W⊗ be a weakly bi-enriched ∞-category that admits left and right tensors and small
conical colimits. The induced functor
EnrFunV,W (N, D) → EnrFunV,W (M, D)
admits a left adjoint ψ! , which is fully faithful if ψ is an embedding.
(2) Let V⊗ → Ass be a presentably monoidal ∞-category, W⊗ → Ass a small ∞-operad and ψ ∶
M⊛ → N⊛ a V, W-enriched functor from a small to a locally small left enriched ∞-category.
Let D⊛ → V⊗ × W⊗ be a left enriched ∞-category that admits left and right tensors and small
conical colimits. The induced functor
(3) Let V⊗ → Ass, W⊗ → Ass be presentably monoidal ∞-categories and ψ ∶ M⊛ → N⊛ a V, W-
enriched functor from a small to a locally small bi-enriched ∞-category. Let D⊛ → V⊗ × W⊗
be a bi-enriched ∞-category that admits left and right tensors and small conical colimits. The
induced functor
(4) For every V, W-enriched functor F ∶ M⊛ → D⊛ and X ∈ N in all cases there is an equivalence
n m
ψ! (F)(X) ≃ colimV1 ,...,Vn ,ψ(Y),W1,...,Wm→X ⊗ Vi ⊗ F(Y) ⊗ ⊗ Wj .
i=1 j=1
Proof. (1): By Corollary 4.38 the weakly bi-enriched ∞-category D → PEnv(V)⊗ × PEnv(W)⊗ is
⊛
the pullback of a bitensored ∞-category D̄⊛ → V⊗ × W⊗ compatible with small colimits. We apply
Proposition 2.62.
(2): The presentably monoidal ∞-category V⊗ → Ass is κ-compactly generated for some small
regular cardinal κ. Let ψκ ∶ M⊛ κ → Nκ be the pullback of ψ along (V ) ⊂ V , which is a V , W-
⊛ κ ⊗ ⊗ κ
enriched functor between totally small weakly bi-enriched ∞-categories. By Corollary 4.34 the induced
functors
EnrFunV,W (N, D) → EnrFunVκ ,W (Nκ , Dκ ),
EnrFunV,W (M, D) → EnrFunVκ ,W (Mκ , Dκ )
are equivalences. Hence (2) follows from (1), the statement that the induced functor
EnrFunVκ ,W (Nκ , Dκ ) → EnrFunVκ ,W (Mκ , Dκ )
admits a left adjoint, which is fully faithful if ψκ is an embedding.
(3): The proof of (3) is similar: the presentably monoidal ∞-categories V⊗ → Ass, W⊗ → Ass are
κ-compactly generated, τ -compactly generated, respectively, for some small regular cardinals κ, τ . Let
ψκ,τ ∶ M⊛
κ,τ → Nκ,τ be the pullback of ψ along (V ) × (W ) ⊂ V × W , which is a V , W -enriched
⊛ κ ⊗ τ ⊗ ⊗ ⊗ κ τ
functor between totally small weakly bi-enriched ∞-categories. By Corollary 4.34 the induced functors
EnrFunV,W (N, D) → EnrFunVκ ,Wτ (Nκ,τ , Dκ,τ ),
EnrFunV,W (M, D) → EnrFunVκ ,Wτ (Mκ,τ , Dκ,τ )
are equivalences. Hence (3) follows from (1), the statement that the induced functor
EnrFunVκ ,Wτ (Nκ,τ , Dκ,τ ) → EnrFunVκ ,Wτ (Mκ,τ , Dκ,τ )
admits a left adjoint, which is fully faithful if ψκ,τ is an embedding. (4) holds by Proposition 2.62.

Notation 5.17. Let M⊛ → V⊗ , M′⊛ → V⊗ be weakly left enriched ∞-categories, N⊛ → V⊗ ×W⊗ , N′⊛ →
V⊗ ×W⊗ be weakly bi-enriched ∞-categories, F ∶ M′⊛ → M⊛ a left V-enriched functor and G ∶ N⊛ → N′⊛
a V, W-enriched functor. Let
EnrFunV,∅ (F, G) ∶ EnrFunV,∅ (M, N)⊛ → EnrFunV,∅ (M′ , N′ )⊛
be the right W-enriched functor corresponding to the V, W-enriched functor
F×EnrFunV,∅ (M,N)⊛
M′⊛ × EnrFunV,∅ (M, N)⊛ ÐÐÐÐÐÐÐÐÐÐÐÐ→ M⊛ × EnrFunV,∅ (M, N)⊛ → N⊛ Ð
→ N′⊛ .
G
Remark 5.18. The right W-enriched functor

EnrFunV,∅ (F, G) ∶ EnrFunV,∅ (M, N)⊛ → EnrFunV,∅ (M′ , N′ )⊛
induces on underlying ∞-categories the canonical functor EnrFunV,∅ (M, N) → EnrFunV′ ,∅ (M′ , N′ ).
Remark 5.19. If N⊛ → V⊗ × W⊗ , N′⊛ → V⊗ × W⊗ admit right tensors and G ∶ N⊛ → N′⊛ is a right W-
linear functor, then EnrFunV,∅ (F, G) ∶ EnrFunV,∅ (M, N)⊛ → EnrFunV,∅ (M′ , N′ )⊛ is a right W-linear
functor between weakly bi-enriched ∞-categories that admit right tensors by Corollary 5.15.
Proposition 5.20. (1) Let M⊛ → V⊗ , M′⊛ → V⊗ be totally small weakly left enriched ∞-categories,
F ∶ M → M a left V-enriched functor and N⊛ → V⊗ × W⊗ a weakly bi-enriched ∞-category
′⊛ ⊛
that admits small conical colimits and left and right tensors. The right W-enriched functor
EnrFunV,∅ (F, N) ∶ EnrFunV,∅ (M, N)⊛ → EnrFunV,∅ (M′ , N)⊛
admits a right W-enriched left adjoint
F! ∶ EnrFunV,∅ (M′ , N)⊛ → EnrFunV,∅ (M, N)⊛ .
(2) Let M⊛ → V⊗ , M′⊛ → V⊗ be small left enriched ∞-categories, F ∶ M′⊛ → M⊛ a left V-enriched
functor and N⊛ → V⊗ × W⊗ a weakly bi-enriched ∞-category that admits small conical colimits
and left and right tensors whose underlying weakly left enriched ∞-category is a left enriched
∞-category. The right W-enriched functor
EnrFunV,∅ (F, N) ∶ EnrFunV,∅ (M, N)⊛ → EnrFunV,∅ (M′ , N)⊛
admits a right W-enriched left adjoint
F! ∶ EnrFunV,∅ (M′ , N)⊛ → EnrFunV,∅ (M, N)⊛ .
Proof. The induced functor F∗ ∶ EnrFunV,∅ (M, N) → EnrFunV,∅ (M′ , N) admits a left adjoint F! under
the assumptions of (1) by Proposition 5.16 (1) and under the assumptions of (2) by Proposition 5.16
(2). By Remark 5.19 the right W-enriched functor EnrFunV,∅ (M, N)⊛ → EnrFunV,∅ (M′ , N)⊛ induced
by F preserves right tensors and source and target of it admit right tensors since N⊛ → V⊗ ×W⊗ admits
right tensors. Using Definition 2.68 (2) it is enough to prove that for every m ≥ 0 and W1 , ..., Wm ∈ W
and left V-linear functor H ∶ M⊛ → N⊛ the following canonical morphism is an equivalence:
(F! ○ ((−) ⊗ Wm ) ○ ... ○ ((−) ⊗ W1 ))(H) → (((−) ⊗ Wm ) ○ ... ○ ((−) ⊗ W1 ) ○ F! )(H).
By Proposition 2.62 (1) the latter morphism induces at X ∈ M the canonical morphism
n
colimV1 ,...,Vn ,F(Y)→X(⊗ Vi ⊗ (((−) ⊗ Wm ) ○ ... ○ ((−) ⊗ W1 ) ○ H(Y)))
i=1
n
→ ((−) ⊗ Wm ) ○ ... ○ ((−) ⊗ W1 ) ○ (colimV1 ,...,Vn ,F(Y)→X ⊗ Vi ⊗ H(Y)).
i=1
The latter morphism is an equivalence since the functors (−) ⊗ Wi ∶ N → N preserve small conical
colimits for 1 ≤ i ≤ m.

5.2. Enriched Yoneda-lemma. In this subsection we prove an enriched Yoneda-lemma.
Lemma 5.21. Let B be a locally small ∞-category and C an ∞-category that admits small colimits.
(1) For every Z ∈ B the functor evBZ ∶ Fun(B, C) → C evaluating at Z ∈ B admits a left adjoint that
sends Y ∈ C to the functor B(Z, −) ⊗ Y ∶ B → C.
(2) Let B be a small space. For every functor H ∶ B → C there is a canonical equivalence H ≃
colimZ∈B B(Z, −) ⊗ H(Z) in Fun(B, C).
(3) Let B be a small ∞-category. The functor G ∶ Fun(B, C) → Fun(B≃ , C) restricting along the
inclusion B≃ ⊂ B admits a left adjoint and is monadic. The left adjoint sends the functor
B≃ (Z, −) ⊗ Y ∶ B≃ → C for Y ∈ C, Z ∈ B to the functor B(Z, −) ⊗ Y ∶ B → C.
(4) Let B be a small ∞-category. The ∞-category Fun(B, C) is generated under small colimits by
{B(Z, −) ⊗ Y ∶ B → C ∣ Z ∈ B, Y ∈ C}.
Proof. (1): By the Yoneda-lemma for ∞-categories there is a canonical equivalence
Z ○ C(Y, −)∗ ≃ C(Y, −) ○ evZ .
Fun(B, C)(B(Z, −) ⊗ Y, −) ≃ Fun(B, S)(B(Z, −), −) ○ C(Y, −)∗ ≃ evB B
(2) Let B be a locally small space. Then for all functors H, H′ ∶ B → C the canonical equivalence
Fun(B, C)(colimZ∈B B(Z, −) ⊗ H(Z), H′ ) ≃ lim Fun(B, C)(B(Z, −) ⊗ H(Z), H′ )
Z∈B
≃ lim C(H(Z), H (Z)) ≃ Fun(B, C)(H, H′ )

′
Z∈B
represents an equivalence H ≃ colimZ∈B B(Z, −) ⊗ H(Z) in Fun(B, C).
(3): Let B be a small ∞-category. To see that G admits a left adjoint, we need to show that for
overy H ∈ Fun(B≃ , C) the functor Fun(B≃ , C)(H, −) ○ G ∶ Fun(B, C) → S is corepresentable. The full
subcategory of Fun(B≃ , C) spanned by the functors H such that the functor Fun(B≃ , C)(H, −) ○ G ∶
Fun(B, C) → S is corepresentable, is closed under small colimits because B is small and C and so
Fun(B, C) admit small colimits. By the first part of the proof the ∞-category Fun(B≃ , C) is generated
under small colimits by the functors B≃ (Z, −) ⊗ Y ∶ B≃ → C for Y ∈ C, Z ∈ B. Therefore G admits a
left adjoint if for every Y ∈ C, Z ∈ B the functor Fun(B≃ , C)(B≃ (Z, −) ⊗ Y, −) ○ G ∶ Fun(B≃ , C) → S is
corepresentable. There is a canonical equivalence, which implies (2):
Fun(B≃ , C)(B≃ (Z, −) ⊗ Y, −) ○ G ≃ C(Y, −) ○ evB
Z ○ G ≃ C(Y, −) ○ evZ ≃ Fun(B, C)(B(Z, −) ⊗ Y, −).
≃
B
(4): The functor G is conservative because the inclusion B≃ ⊂ B is essentially surjective. The
functor G preserves small colimits since such a formed object-wise. Thus G is monadic [16, Theorem
4.7.3.5]. This guarantees that Fun(B, C) is generated under small colimits by the essential image
of F. By the first part of the proof Fun(B≃ , C) is generated under small colimits by the functors
B≃ (Z, −) ⊗ Y ∶ B≃ → C for Y ∈ C, Z ∈ B so that Fun(B, C) is generated under small colimits by the
images under F of all functors B≃ (Z, −) ⊗ Y ∶ B≃ → C for Y ∈ C, Z ∈ B.

Theorem 5.22. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category and N⊛ → V⊗ × W⊗ a weakly
bi-enriched ∞-category that admits left and right tensors and small conical colimits.
(1) For every H ∈ EnrFunV,W (M, N), X ∈ M, N ∈ N and V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0
the following map is an equivalence:
EnrFunV,W (M, N)(LMorM̄ (X, −) ⊗ N, H) → N(N, H(X)).
(2) If M is small, then EnrFunV,W (M, N) is generated under small colimits by {LMorM̄ (X, −)⊗N ∣
X ∈ M, N ∈ N} and the forgetful functor EnrFunV,W (M, N) → Fun(M, N) admits a left adjoint
and is monadic. The left adjoint sends M(Z, −)⊗N to LMorM̄ (Z, −)⊗N for every Z ∈ M, N ∈ N.
(3) If M is small, the forgetful functor ν ∶ EnrFunV,W (M, N) → Fun(M≃ , N) admits a left adjoint
φ and is monadic and for every F ∶ M≃ → N the canonical morphism
colimZ∈M≃ (LMorM̄ (Z, −) ⊗ F(Z)) → φ(F)
in EnrFunV,W (M, N) is an equivalence.
Proof. (1): Let N̄⊛ ⊂ PBEnv(N)⊛ be the full weakly bi-enriched subcategory spanned by N. By
Lemma 4.37 (5) the bi-enriched ∞-category N̄⊛ → PEnv(V)⊗ × PEnv(W)⊗ is a bitensored ∞-category
compatible with small colimits. Consider the PEnv(V), PEnv(W)-enriched adjunction
(−) ⊗ X ⊗ (−) ∶ (PEnv(V) ⊗ PEnv(W))⊛ ⇄ PBEnv(M)⊛ ∶ LMorPBEnv(M) (X, −).
By Proposition 4.46 the right adjoint preserves small colimits and is PEnv(V), PEnv(W)-linear. Hence
by Remark 2.73 the latter PEnv(V), PEnv(W)-enriched adjunction induces an adjunction
LMorPBEnv(M) (X, −)∗ ∶ LinFunLPEnv(V),PEnv(W) (PBEnv(M), N̄) ⇄
LinFunLPEnv(V),PEnv(W) (PEnv(V) ⊗ PEnv(W), N̄) ∶ ((−) ⊗ X)∗ .

Consider the commutative square
LinFunLPEnv(V),PEnv(W) (PBEnv(M), N̄)

≃ / EnrFunV,W (M, N)
((−)⊗X)∗ evX

LinFunLPEnv(V),PEnv(W) (PEnv(V) ⊗ PEnv(W), N̄)
≃

/ N.
By Proposition 2.109 and Proposition 2.51 the horizontal functors are equivalences. Consequently,
the right vertical functor admits a left adjoint that sends N ∈ N to the following composition of
V, W-enriched functors:
M⊛ ⊂ PBEnv(M)⊛ ÐÐÐÐÐÐÐÐÐÐÐ→ (PEnv(V) ⊗ PEnv(W))⊛ ÐÐÐÐÐÐ→ N⊛ .

LMorPBEnv(M) (X,−) (−)⊗N⊗(−)
The unit of this adjunction is induced by the canonical morphism 1 → LMorPBEnv(M) (X, X) in
PEnv(V) ⊗ PEnv(W). This proves (1).
(2): Let θ ∶ EnrFunV,W (M, N) → Fun(M, N) be the forgetful functor. We first show that θ admits
a left adjoint. Let Θ ⊂ Fun(M, N) be the full subcategory spanned by those H such that the functor
Fun(M, N)(H, −) ○ θ ∶ EnrFunV,W (M, N) → S is corepresentable. The functor θ admits a left adjoint
if and only if Θ = Fun(M, N). By Lemma 2.52 (1) the ∞-category EnrFunV,W (M, N) admits small
colimits preserved by θ. By This guarantees that Θ is closed under small colimits. By Lemma 5.21 (4)
the ∞-category Fun(M, N) is generated under small colimits by {M(Z, −)⊗N ∶ M → N ∣ N ∈ N, Z ∈ M}.
Therefore it is enough to see that for every N ∈ N, Z ∈ M the functor Fun(M, N)(M(Z, −) ⊗ N, −) ○ θ ∶
EnrFunV,∅ (M, N) → S is corepresentable. By Lemma 5.21 (1) for any Z ∈ M the functor N →
Fun(M, N), N ↦ M(Z, −) ⊗ N is left adjoint to the functor evZ evaluating at Z. Thus by (1) the
functor Fun(M, N)(M(Z, −) ⊗ N, −) ○ θ ≃ N(N, −) ○ evZ ○ θ ∶ EnrFunV,W (M, N) → S is equivalent to
the corepresentable functor EnrFunV,W (M, N)(LMorM̄ (Z, −) ⊗ N, −) ∶ EnrFunV,W (M, N) → S. So θ
admits a left adjoint that sends M(Z, −) ⊗ N to LMorM̄ (Z, −) ⊗ N for every Z ∈ M, N ∈ N.
By [16, 4.7.3.5.] this guarantees that θ is monadic so that EnrFunV,W (M, N) is generated under
small colimits by the essential image of the left adjoint of θ. Since the ∞-category Fun(M, N) is
generated under small colimits by the functors M(Z, −) ⊗ N ∶ M → N for Z ∈ M, N ∈ N, the ∞-category
EnrFunV,W (M, N) is generated under small colimits by the objects LMorM̄ (Z, −)⊗N for Z ∈ M, N ∈ N.
(3): By (2), Lemma 5.21 (3) and [16, 4.7.3.5.] the forgetful functor ν ∶ EnrFunV,W (M, N) →
Fun(M≃ , N) has a left adjoint φ and is monadic. The left adjoint sends M≃ (Z, −)⊗N to LMorM̄ (Z, −)⊗
N for every Z ∈ M, N ∈ N.
For every functor F ∶ M≃ → N the unit F → ν(φ(F)) in Fun(M≃ , N) gives rise to a map
ρ ∶ colimZ∈M≃ (LMorM̄ (Z, −) ⊗ F(Z)) → φ(F)
in EnrFunV,W (M, N) whose summand LMorM̄ (Z, −) ⊗ F(Z)) → φ(F) for Z ∈ M≃ corresponds to the
morphism F(Z) → φ(F)(Z) in N. Since ν, φ preserve small colimits and N admits small conical colimits,
the full subcategory of Fun(M≃ , N) spanned by the functors F, for which the ρ is an equivalence is
closed under small colimits. So by Lemma 5.21 (4) we can assume that F = M≃ (Z, −) ⊗ N ∶ M≃ → N
for N ∈ N and Z ∈ M. By Lemma 5.21 (2) there is a canonical equivalence
F ≃ colimZ∈M≃ M≃ (Z, −) ⊗ F(Z)
in Fun(M≃ , N). Note that ν ○ LMorN̄ (F(Z), −)∗ ≃ LMorN̄ (F(Z), −)∗ ○ ν so that by adjointness
φ ○ ((−) ⊗ F(Z))∗ ≃ ((−) ⊗ F(Z))∗ ○ φ.
Hence there is a canonical equivalence
colimZ∈M≃ LMorM̄ (Z, −) ⊗ F(Z) ≃ colimZ∈M≃ φ(M≃ (Z, −)) ⊗ F(Z)
≃ colimZ∈M≃ φ(M≃ (Z, −) ⊗ F(Z)) ≃ φ(colimZ∈M≃ M≃ (Z, −) ⊗ F(Z)) ≃ ν(φ(F))
in EnrFunV,W (M, N), which canonically identifies with ρ.

Corollary 5.23. Let M⊛ → V⊗ be a weakly left enriched ∞-category and N⊛ → V⊗ × W⊗ a weakly
bi-enriched ∞-category that admits left tensors.
(1) For every H ∈ EnrFunV,∅ (M, N) and X ∈ M, N ∈ N and W1 , ..., Wm ∈ W for m ≥ 0 the following
canonical map is an equivalence:
MulEnrFunV,∅ (M,N) (LMorM̄ (X, −) ⊗ N, W1 , ..., Wm ; H) → MulN (N, W1 , ..., Wm ; H(X)).
(2) If M is small and N⊛ → V⊗ × W⊗ admits small conical colimits, the right W-enriched forgetful
functor EnrFunV,∅ (M, N)⊛ → Fun(M, N)⊛ has a right W-enriched left adjoint that sends
M(Z, −) ⊗ N to LMorM̄ (Z, −) ⊗ N for every Z ∈ M, N ∈ N.
(3) If M is small and N⊛ → V⊗ × W⊗ admits small conical colimits, the right W-enriched forgetful
functor ν ∶ EnrFunV,∅ (M, N)⊛ → Fun(M≃ , N)⊛ admits a right W-enriched left adjoint φ and
for every F ∶ M≃ → N the following canonical morphism in EnrFunV,∅ (M, N) is an equivalence:
colimZ∈M≃ (LMorM̄ (Z, −) ⊗ F(Z)) → φ(F).
Proof. Statement (1) is equivalent to say that for every X ∈ M the right W-enriched functor γ ∶
EnrFunV,∅ (M, N)⊛ → N⊛ evaluating at X admits a right W-enriched left adjoint and the unit of
the adjunction is induced by the canonical morphism 1 → LMorM̄ (X, X) in PEnv(V). By Theorem
5.22 (1) the functor EnrFunV,∅ (M, N) → N evaluating at X admits a left adjoint that sends N ∈ N
to LMorM̄ (X, −) ⊗ N, and the unit of this adjunction is induced by the canonical morphism 1 →
LMorM̄ (X, X) in PEnv(V). By Proposition 5.15 the right W-enriched functor γ is a right W-linear
functor between weakly right W-enriched ∞-categories that admit right tensors. Therefore by Lemma
2.69 the result follows from the fact that for every N ∈ N, W ∈ W the morphism
κ ∶ LMorM̄ (X, −) ⊗ (N ⊗ W) → (LMorM̄ (X, −) ⊗ N) ⊗ W
corresponding to the morphism 1 ⊗ N ⊗ W → LMorM̄ (X, X) ⊗ N ⊗ W is the canonical equivalence.

(2): By Proposition 5.15 the right W-enriched forgetful functor EnrFunV,∅ (M, N)⊛ → Fun(M, N)⊛
is a right W-linear functor between weakly right W-enriched ∞-categories that admit right tensors.
By Theorem 5.22 (2) the forgetful functor EnrFunV,∅ (M, N) → Fun(M, N) admits a left adjoint φ.
Thus by Lemma 2.69 it is enough to see that for every H ∈ Fun(M, N), W ∈ W the canonical morphism
λ ∶ φ(H ⊗ W) → φ(H) ⊗ W is an equivalence. By Lemma 5.21 (2) the ∞-category Fun(M, N) is
generated under small colimits by {M(X, −) ⊗ N ∶ M → N ∣ N ∈ N, X ∈ M}. By Proposition 5.15
forming right tensors in Fun(M, N) and EnrFunV,∅ (M, N) preserves small colimits. So it suffices to
show that λ is an equivalence for H = M(X, −) ⊗ N ∶ M → N for any N ∈ N, X ∈ M. In this case λ
identifies with κ. The proof of (3) is similar to the one of (2), where we use Theorem 5.22 (3).

Corollary 5.24. Let V⊗ → Ass be a monoidal ∞-category compatible with small colimits and M⊛ → V⊗
a small left enriched ∞-category.
(1) Then EnrFunV,∅ (M, V) is generated under small colimits and left tensors by {LMorM̄ (X, −) ∣
X ∈ M}.
(2) For every X ∈ M the right V-enriched functor
RMorEnrFunV,∅ (M,V) (LMorM (X, −), −) ∶ EnrFunV,∅ (M, V)⊛ → V⊛
is equivalent to the functor evaluating at X and so preserves small colimits and is right V-
linear.
Remark 5.25. Let M⊛ → V⊗ , N⊛ → V⊗ be left enriched ∞-categories, F ∶ M⊛ → N⊛ a left V-enriched
functor and X ∈ M. The morphism 1V → LMorN (F(X), F(X)) in V corresponds to a morphism
LMorM (X, −) → LMorN (F(X), −) ○ F in EnrFunV,∅ (M, V) whose component at any Y ∈ M is the
canonical morphism LMorM (X, Y) → LMorN (F(X), F(Y)) in V.
Proposition 5.26. Let M⊛ → V⊗ × W⊗ be a small weakly bi-enriched ∞-category, N⊛ → V⊗ ×
W⊗ a weakly bi-enriched ∞-category that admits left and right tensors and small conical colim-
its. Every V, W-enriched functor F ∶ M⊛ → N⊛ is the colimit of a canonical simplicial object X in
EnrFunV,W (M, N) such that for every n ≥ 0 there is a canonical equivalence in EnrFunV,W (M, N) ∶
Xn ≃ colimZ1 ,....,Zn ∈M≃ LMorM̄ (Zn , −)⊗LMorM̄ (Zn−1 , Zn )⊗...⊗LMorM̄ (Z2 , Z3 )⊗LMorM̄ (Z1 , Z2 )⊗F(Z1 ).
Proof. By Theorem 5.22 (3) the forgetful functor ν ∶ EnrFunV,W (M, N) → Fun(M≃ , N) admits a left
adjoint φ and is monadic. This guarantees that F admits a monadic resolution, i.e. F is the colimit
of a canonical simplicial object X in EnrFunV,W (M, N) such that for every n ≥ 0 there is a canonical
equivalence in EnrFunV,W (M, N) ∶
Xn ≃ (φ ○ ν)○n (F).
By Theorem 5.22 (3) and induction over n ≥ 0 there is a canonical equivalence in EnrFunV,W (M, N) ∶
Xn ≃ (φ ○ ν)○n (F) ≃
colimZ1 ,....,Zn ∈M≃ LMorM̄ (Zn , −) ⊗ LMorM̄ (Zn−1 , Zn ) ⊗ ... ⊗ LMorM̄ (Z2 , Z3 ) ⊗ LMorM̄ (Z1 , Z2 ) ⊗ F(Z1 ).

Corollary 5.27. Let M⊛ → V⊗ × W⊗ be a small weakly bi-enriched ∞-category and N⊛ → V⊗ × W⊗
a weakly bi-enriched ∞-category that admits left and right tensors. For every V, W-enriched functors
F, G ∶ M⊛ → N⊛ there is a canonical equivalence of spaces
EnrFunV,W (M, N)(F, G) ≃ limop lim N(LMorM̄ (Zn−1 , Zn )⊗...⊗LMorM̄ (Z1 , Z2 )⊗F(Z1 ), G(Zn )).
[n]∈∆ Z1 ,....,Zn ∈M≃
Corollary 5.23 (1) implies the following:

Corollary 5.28. Let M⊛ → V⊗ × W⊗ be a small weakly bi-enriched ∞-category and N⊛ → V⊗ × W⊗ a

weakly bi-enriched ∞-category that admits left tensors. For every left V-enriched functors F, G ∶ M⊛ →
N⊛ there is a canonical equivalence in W ∶
RMorEnrFunV,∅ (M,N) (F, G) ≃
lim lim RMorN̄ (LMorM̄ (Zn−1 , Zn ) ⊗ ... ⊗ LMorM̄ (Z1 , Z2 ) ⊗ F(Z1 ), G(Zn )).
[n]∈∆op Z1 ,....,Zn ∈M≃
5.3. Enriched Yoneda-embedding and opposite enriched ∞-category. Next we use the en-
riched Yoneda-lemma to construct an opposite enriched ∞-category and an enriched Yoneda-embedding.
The next definition is [10, Notation 9.2.]:
Definition 5.29. Let M⊛ → V⊗ be a left enriched ∞-category. The opposite right V-enriched ∞-
category of M⊛ → V⊗ is the full weakly right V-enriched subcategory
(Mop )⊛ ⊂ EnrFunV,∅ (M, V)⊛
spanned by all left V-enriched functors MorM (X, −) ∶ M⊛ → V⊛ for X ∈ M.
The following remark, proposition and lemma justify the terminology of Definition 5.29:
Remark 5.30. By construction 3.6 and Corollary 4.49 for every weakly left enriched ∞-category
M⊛ → V⊗ there is an embedding ξ ∶ Mop → EnrFunV,∅ (M, PEnv(V))⊛ sending X ∈ M to LMorM̄ (X, −).
If M⊛ → V⊗ is a left enriched ∞-category, by Remark 3.20 the embedding ξ induces an embedding
Mop → EnrFunV,∅ (M, V)⊛ ⊂ EnrFunV,∅ (M, PEnv(V))⊛ denoted by the same name, which sends X ∈ M
to LMorM̄ (X, −). The embedding ξ induces an equivalence Mop ≃ (Mop )⊛ ×V⊗ ∅⊗ .
Proposition 5.31. Let M⊛ → V⊗ be a left enriched ∞-category. The weakly right enriched ∞-category
(Mop )⊛ → V⊗ exhibits Mop as right enriched in V, where for any X, Y ∈ M there is a canonical
equivalence
RMorMop (X, Y) ≃ LMorM (Y, X).
Proof. By Corollary 5.23 for any X, Y ∈ M and V1 , ..., Vn ∈ V for n ≥ 0 the canonical map
MulMop (V1 , ..., Vn , LMorM (X, −); LMorM (Y, −)) → MulV (V1 , ..., Vn , LMorM (Y, X))
is an equivalence. So the result follows.
Notation 5.32. Let V⊗ → Ass, W⊗ → Ass be small ∞-operads. Via the canonical equivalences
V ωBEnrW ≃ P(Env(W)rev ×Env(V)) LEnr, W ωBEnrV ≃ REnrP(Env(W)rev ×Env(V))
of Corollary 4.32 we define for every weakly bi-enriched ∞-category M⊛ → V⊗ ×W⊗ an opposite weakly
bi-enriched ∞-category (Mop )⊛ → W⊗ × V⊗ extending Definition 5.29.
Remark 5.33. Let M⊛ → V⊗ , N⊗ → W⊗ be left enriched ∞-categories. The V × W, V × W-enriched
functor
(M⊛ ×Ass N⊛ ) × (EnrFunV,∅ (M, V)⊛ ×Ass EnrFunV,∅ (N, V)⊛ ) ≃
(M⊛ × EnrFunV,∅ (M, V)⊛ ) ×Ass×Ass (N⊛ × EnrFunV,∅ (N, V)⊛ ) → V⊛ ×Ass×Ass W⊛ ,
which is the product of the V, V-enriched and W, W-enriched evaluation functors, corresponds to a
right V × W-enriched functor
ρ ∶ EnrFunV,∅ (M, V)⊛ ×Ass EnrFunV,∅ (N, V)⊛ → EnrFunV×W,∅ (M × N, V × W)⊛ .
The latter restricts to a right V × W-enriched equivalence
ρ′ ∶ (Mop )⊛ ×Ass (Nop )⊛ ≃ ((M × N)op )⊛ .
This follows since ρ sends LMorM (X, −), LMorN (Y, −) for X ∈ M, Y ∈ N to LMorM (X, −)×LMorN (Y, −) ≃
LMorM×N ((X, Y), −) and so restricts to ρ′ and via Corollary 5.24 the enriched functor ρ′ induces on
right morphism objects the following equivalence for X′ ∈ M, Y′ ∈ N:
LMorM (X′ , X) × LMorN (Y′ , Y) ≃
RMorMop (LMorM (X, −), LMorM (X′ , −)) × RMorNop (LMorN (Y, −), LMorN (Y′ , −)) →
RMor(M×N)op (LMorM×N ((X, Y), −), LMorM×N ((X′ , Y′ ), −)) ≃ LMorM×N ((X′ , Y′ ), (X, Y)).
Proposition 5.34. Let M⊛ → V⊗ be a left enriched ∞-category, O⊛ → W⊗ a right enriched ∞-category
and N⊛ → V × W⊗ a bi-enriched ∞-category.
(1) There is a W, V-enriched equivalence
((M × O)op )⊛ ≃ (Mop )⊛ × (Oop )⊛ .
(2) There is a right W-enriched equivalence
EnrFunV,∅ (M, N) ≃ EnrFun∅,V (Mop , Nop )op .
Proof. (1): By Remark 5.33 there is a left V, Wrev -enriched equivalence
(M × Orev )op )⊛ ≃ (Mop )⊛ ×Ass ((Oop )rev )⊛ .
Under the equivalence V BPEnrW ≃ V×Wrev LPEnr the latter equivalence corresponds to the V, W-
enriched equivalence ((M × O)op )⊛ ≃ (Mop )⊛ × (Oop )⊛ .
(2): The claimed equivalence is represented by the following chain of natural equivalences:
op op op
∅ ωBEnrW (O, EnrFun∅,V (M , N ) ) ≃ W ωBEnr∅ (Oop , EnrFun∅,V (Mop , Nop ))
≃ W ωBEnrV (Oop × Mop , Nop )≃ W ωBEnrV ((M × O)op , Nop ) ≃
V ωBEnrW (M × O, N) ≃ ∅ ωBEnrW (O, EnrFunV,∅ (M, N)).

Lemma 5.35. Let V → Ass be a presentably monoidal ∞-category, M → V , N → V small left
⊗ ⊛ ⊗ ⊛ ⊗
enriched ∞-categories and F ∶ M⊛ → N⊛ a left V-enriched functor. There is a commutative square:

Fop
Mop / Nop
ξM ξN
F!

EnrFunV,∅ (M, V) / EnrFunV,∅ (N, V).
Proof. The enriched adjunction F! ∶ PLEnv(M)⊛ ⇄ PLEnv(N)⊛ ∶ F∗ gives rise to an adjunction

(F∗ )∗ ∶ EnrFunR
PEnv(V),∅ (PLEnv(M), PEnv(V)) ⇄ EnrFunPEnv(V),∅ (PLEnv(N), PEnv(V)) ∶ (F! ) .
R ∗
There is a commutative diagram

Fop
Mop / Nop

PLEnv(M)op / PLEnv(N)op
≃ ≃

EnrFunLPEnv(V),∅ (PEnv(V), PLEnv(M))op / EnrFunL
PEnv(V),∅ (PEnv(V), PLEnv(N))op
≃ ≃
(F∗ )∗

EnrFunR
PEnv(V),∅ (PLEnv(M), PEnv(V)) / EnrFunR
PEnv(V),∅ (PLEnv(N), PEnv(V))
Let αM be the left vertical functor in the diagram so that αN is the right vertical functor in the
diagram. The diagram provides an equivalence (F∗ )∗ ○ αM ≃ αN ○ Fop . Composing the composition
αM → (F! )∗ ○ (F∗ )∗ ○ αM ≃ (F! )∗ ○ αN ○ Fop
of maps of functors Mop → EnrFunR
PEnv(V),∅ (PLEnv(M), PEnv(V)) with the functor
PEnv(V),∅ (PLEnv(M), PEnv(V)) → EnrFunV,∅ (M, PEnv(V))

EnrFunR
gives a morphism ξM → F∗ ○ ξN ○ Fop of functors Mop → EnrFunV,∅ (M, V) ⊂ EnrFunV,∅ (M, PEnv(V)).
The composition
F! ○ ξM → F! ○ F∗ ○ ξN ○ Fop → ξN ○ Fop
induces at Z ∈ M the morphism κ ∶ F! (LMorM (Z, −)) → LMorN (F(Z), −) in EnrFunV,∅ (N, V) cor-
responding to the morphism 1V → LMorN (F(Z), F(Z)) given by the identity. By Lemma 5.42 the
morphism κ is an equivalence.

Notation 5.36. Let M⊛ → V⊗ be a left enriched ∞-category.
(1) The V, V-enriched left morphism object functor
LMorM ∶ M⊛ × (Mop )⊛ → V⊛
is the restricted V, V-enriched evaluation functor
M⊛ × (Mop )⊛ ⊂ M⊛ × EnrFunV,∅ (M, V)⊛ → V⊛ .
(2) The left V-enriched Yoneda-embedding
ρM ∶ M⊛ → EnrFun∅,V (Mop , V)⊛
is the left V-enriched functor corresponding to the left morphism object functor LMorM ∶
M⊛ × (Mop )⊛ → V⊛ under the equivalence of Proposition 5.11.
Remark 5.37. Notation 5.36 also applies to any right V-enriched ∞-category M⊛ → V⊗ (viewed as a
left Vrev -enriched ∞-category). In this case we obtain a V, V-enriched right morphism object functor
RMorM ∶ (Mop )⊛ × M⊛ → V⊛ ,
which is the restricted V, V-enriched evaluation functor
(Mop )⊛ × M⊛ ⊂ EnrFun∅,V (M, V)⊛ × M⊛ → V⊛ ,
and a corresponding right V-enriched Yoneda-embedding ρM ∶ M⊛ → EnrFunV,∅ (Mop , V)⊛ .
Remark 5.38. The V, V-enriched left morphism object functor LMorM ∶ M⊛ × (Mop )⊛ → V⊛ induces
on underlying weakly left V-enriched ∞-categories the left V-enriched functor ΓM ∶ M⊛ × Mop → V⊛
of Construction 3.6 corresponding to ξ.
Lemma 5.39. Let V⊗ → Ass be a monoidal ∞-category compatible with small colimits, M⊛ →
V⊗ a small left enriched ∞-category and Z ∈ M. By Theorem 5.22 the canonical morphism 1V →
ρ(Z)(ξ(Z)) ≃ LMorM (Z, Z) in V corresponds to a morphism
κ ∶ RMorMop (ξ(Z), −) → ρ(Z)
op
in EnrFun∅,V (M , V). The morphism κ is an equivalence.
Proof. The right V-enriched functor ρ(Z) ∶ (Mop )⊛ → V⊛ induces a morphism
RMorMop (ξ(Z), −) → RMorPEnv(V) (LMorM (Z, Z), −) ○ ρ(Z) → ρ(Z)
in EnrFun∅,V (M , V) whose component at H ∈ Mop is the canonical morphism
op
RMorMop (ξ(Z), H) → RMorPEnv(V) (LMorM (Z, Z), H(Z)) → H(Z)

in V, which is an equivalence by Corollary 5.23.

Proposition 5.40. Let M⊛ → V⊗ be a small left enriched ∞-category. The left V-enriched Yoneda-
embedding ρ ∶ M⊛ → EnrFun∅,V (Mop , V)⊛ induces an equivalence M⊛ → ((Mop )op )⊛ .
Proof. By Lemma 5.39 the left V-enriched Yoneda-embedding ρM ∶ M⊛ → EnrFunV,∅ (Mop , V)⊛ in-
duces a left V-enriched functor M⊛ → ((Mop )op )⊛ whose underlying functor M → (Mop )op is essentially
surjective. By Remark 5.25 the left V-enriched functor ρ ∶ M⊛ → (Mop )op )⊛ induces a morphism
κ ∶ LMorM (Z, −) → LMor(Mop )op (ρ(Z), −) ○ ρ
in EnrFunV,∅ (M, V). By Remark 5.25 the left V-enriched functor evξ(Z) ∶ EnrFun∅,V (Mop , V)⊛ → V⊛
evaluating at ξ(Z) ∈ Mop induces the following morphism λ in EnrFunV,∅ (M, V):
LMor(Mop )op (ρ(Z), −)○ρ → LMorPEnv(V) (LMorM (Z, Z), −)○evξ(Z) ○ρ → evξ(Z) ○ρ ≃ ξ(Z) = LMorM (Z, −),
which is an equivalence by Corollary 5.23. By Corollary 5.23 the composition λ ○ κ is the identity
because λZ ○ κZ sends the identity of Z to the identity of Z. So the result follows.
Remark 5.41. Let M⊛ → V⊗ be a small left enriched ∞-category. By Proposition 5.40 the left
V-enriched Yoneda-embedding ρM ∶ M⊛ → EnrFun∅,V (Mop , V)⊛ induces an equivalence ρ′M ∶ M⊛ →
((Mop )op )⊛ . By definition ρM ∶ M⊛ → EnrFun∅,V (Mop , V)⊛ corresponds to the V, V-enriched left
morphism object functor LMorM ∶ M⊛ × (Mop )⊛ → V⊛ . Thus LMorM ∶ M⊛ × (Mop )⊛ → V⊛ factors as
V, V-enriched functors
M⊛ × (Mop )⊛ ÐÐÐÐÐÐÐ→ EnrFun∅,V (Mop , V)⊛ × (Mop )⊛ → V⊛ ,

ρM ×(Mop )⊛
which factors as
ρ′M ×(Mop )⊛
M⊛ × (Mop )⊛ ÐÐÐÐÐÐÐ→ (Mop )op )⊛ × (Mop )⊛ ÐÐÐÐÐ→ V⊛ .
RMorMop
In particular, ρMop ∶ (Mop )⊛ → EnrFunV,∅ ((Mop )op , V)⊛ ≃ EnrFunV,∅ (M, V)⊛ , where the latter equiv-
alence precomposes with ρ′ , is the tautological embedding (Mop )⊛ ⊂ EnrFunV,∅ (M, V)⊛ .
Lemma 5.42. Let V⊗ → Ass be a monoidal ∞-category compatible with small colimits, M⊛ →
V⊗ , N⊛ → V⊗ small left enriched ∞-categories and F ∶ M⊛ → N⊛ a left V-enriched functor. For
every X ∈ M the morphism
κ ∶ F! (LMorM (X, −)) → LMorN (F(X), −)
in EnrFunV,∅ (N, V) corresponding to the morphism 1V → LMorN (F(X), F(X)) given by the identity
is an equivalence.
Proof. The morphism κ induces for every H ∈ EnrFunV,∅ (N, V) the identity
H(F(X)) ≃ EnrFunV,∅ (N, V)(MorN (F(X), −), H) → EnrFunV,∅ (N, V)(F! (MorM (X, −)), H)
≃ EnrFunV,∅ (M, V)(MorM (X, −), H ○ F) ≃ H(F(X)),
where we use Corollary 5.23.
Corollary 5.43. Let V⊗ → Ass be a monoidal ∞-category compatible with small colimits, M⊛ →
V⊗ , N⊛ → V⊗ small left enriched ∞-categories and F ∶ M⊛ → N⊛ a left V-enriched functor. The right
V-enriched functor
F! ∶ EnrFunV,∅ (M, V)⊛ → EnrFunV,∅ (N, V)⊛
restricts to a right V-enriched functor (Mop )⊛ → (Nop )⊛ , which we denote by Fop .
Construction 5.44. Let V⊗ → Ass be a monoidal ∞-category compatible with small colimits. By
Proposition 5.20 the functor EnrFunV,∅ (−, V)⊛ ∶ (V ωBEnr∅ )op → ∅ ω BEnr ̂ R . By
̂ V lands in ∅ ω BEnr
V
̂ R ≃ (∅ ω BEnr
Corollary 2.72 there is a canonical equivalence ∅ ω BEnr ̂ L )op . Corollary 5.43 implies that
V V
there is a subfunctor ζV ∶ V ωBEnr∅ → ∅ ωBEnrV ⊂ ∅ ω BEnr
̂ V of the composition
⊛
ÐÐÐÐÐÐÐÐÐÐ→ (∅ ωBEnrR
EnrFunV,∅ (−,V) op ̂ L ⊂ ∅ ω BEnr
̂ V,
V ωBEnr∅ V) ≃ ∅ ω BEnrV
where ζ sends M⊛ → V⊗ to (Mop )⊛ → V⊗ .

Proposition 5.45. Let V⊗ → Ass be a monoidal ∞-category compatible with small colimits. There is
a canonical equivalence
id → ζVrev ○ ζV
of endofunctors of V BEnr∅ whose component at M⊛ → V⊗ is ρM ∶ M⊛ → ((Mop )op )⊛ .
Replacing V⊗ → Ass by (Vrev )⊗ → Ass there is also an equivalence of endofunctors of ∅ BEnrV :
id → ζV ○ ζVrev .
Proof. Let M⊛ → S × V⊗ be a map of cocartesian fibrations over S classifying a functor S → V BEnr∅ ⊂
Cat∞/V⊗ . Let
S×V⊗ ×V⊗ (M × V , S × V )
⊗
(Mop )⊛ ⊂ EnrFunSV,∅ (M, V)⊛ ⊂ FunS×V ⊛ ⊗ ⊛
be the full subcategories spanned by the objects of
S×V⊗ ×V⊗ (M × V , S × V )s ≃ FunV⊗ ×V⊗ (Ms × V , V ),

⊗ ⊗
EnrFunV,∅ (Ms , V)⊛ ⊂ FunS×V ⊛ ⊗ ⊛ V ⊛ ⊗ ⊛
s ) ⊂ EnrFunV,∅ (Ms , V) ⊂ FunS×V⊗ ×V⊗ (M × V , S × V )s ≃ FunV⊗ ×V⊗ (Ms × V , V ),

⊗ ⊗
⊛ ⊛ S×V ⊛ ⊗ ⊛ V ⊛ ⊗ ⊛
(Mop
respectively, for some s ∈ S.
S×V⊗ ×V⊗ (M × V , S × V ) → S × V is a map of cartesian fibrations
⊗
By Remark 5.6 the functor FunS×V ⊛ ⊗ ⊛ ⊗
over S. Since the fiber transports restrict, also EnrFunV,∅ (M, V) → S × V⊗ is a map of cartesian
S ⊛
fibrations over S that induces on the fiber over any s ∈ S the weakly right enriched ∞-category
V⊗ ×V⊗ (Ms × V , V ) → V . This implies that EnrFunV,∅ (M, V) → V is also a map of cocarte-
⊗
S
FunV ⊛ ⊗ ⊛ ⊗ ⊛ ⊗
sian fibrations relative to the collection of cocartesian lifts of inert morphism of Ass preserving the
minimum, whose fiber transports preserve cartesian morphisms over S being projections. This guar-
antees that the map EnrFunSV,∅ (M, V)⊛ → S × V⊗ of cartesian fibrations over S classifies a functor
α ∶ Sop → ∅ BEnr
̂ V . Proposition 5.20 guarantees that α lands in ∅ BEnr ̂ R . This implies by [17, Corol-
V
lary 5.2.2.5.] that EnrFunSV,∅ (M, V)⊛ → S × V⊗ is also a map of cocartesian fibrations over S whose
fiber transports preserve cocartesian lifts of inert morphisms preserving the minimum, and so classifies
̂ L , which by construction factors as S Ð
a functor β ∶ S → ∅ BEnr Ð→ (∅ BEnr
αop
̂ L ⊂ ∅ BEnr
̂ R )op ≃ ∅ BEnr ̂ V.
V V V
By Corollary 5.43 the fiber transports of the map EnrFunSV,∅ (M, V)⊛ → S × V⊗ of cocartesian
fibrations over S restrict to (Mop )⊛ so that also (Mop )⊛ → S × V⊗ is a map of cocartesian fibrations
over S classifying a functor S → ∅ BEnr
̂ V.
The restricted evaluation functor M⊛ ×S (Mop )⊛ ⊂ M⊛ ×S EnrFunSV,∅ (M, V)⊛ → V⊛ over S × V⊗ × V⊗
corresponds to a functor σ ∶ M⊛ → EnrFunS∅,V (Mop , V)⊛ over S × V⊗ that induces on the fiber over any
s ∈ S the embedding ρ ∶ M⊛ op op ⊛ op ⊛
s ≃ ((Ms ) ) ⊂ EnrFun∅,V (Ms , V) of Proposition 5.40. Lemma 5.42
guarantees that σ is a map of cocartesian fibrations over S. Hence σ is an embedding whose essential
image is ((Mop )op )⊛ . Thus σ induces an equivalence σ ′ ∶ M⊛ → ((Mop )op )⊛ that induces on the fiber
over any s ∈ S the equivalence ρ ∶ M⊛ op op ⊛
s ≃ ((Ms ) ) .
If M → S × V classifies the identity of V BEnr∅ , the map EnrFunSV,∅ (M, V)⊛ → S × V⊗ of cartesian
⊛ ⊗
fibrations over S classifies the functor

EnrFunV,∅ (−, V)⊛ ∶ (V ωBEnr∅ )op → ∅ ω BEnr
̂R ̂
V ⊂ ∅ BEnrV
of Construction 5.44 by [4, Proposition 7.3.]. Thus the map EnrFunSV,∅ (M, V)⊛ → S×V⊗ of cocartesian
fibrations over S classifies the composition
(−,V)⊛
ÐÐÐÐÐÐÐÐÐÐ→ (∅ ωBEnrR
EnrFun
V,∅ op ̂ L ⊂ ∅ ω BEnr
̂V
V ωBEnr∅ V) ≃ ∅ ω BEnrV
and the map (Mop )⊛ → S × V⊗ of cocartesian fibrations over S classifies the functor ζV ∶ V BEnr∅ →
∅ BEnrV of Construction 5.44.
Sending a map M⊛ → S × V⊗ of cocartesian fibrations over S classifying a functor S → V BEnr∅ to
the functor S → ∅ BEnrV classified by the map (Mop )⊛ → S × V⊗ of cocartesian fibrations over S gives
a natural map
∞ )(S, V BEnr∅ ) → Ho(Cat∞ )(S, ∅ BEnrV ).
̂
Ho(Cat ̂
By the classical Yoneda-lemma this map postcomposes with the functor ζV ∶ V BEnr∅ → ∅ BEnrV .
In particular for any map M⊛ → S × V⊗ of cocartesian fibrations over S classifying a functor τ ∶
S → V BEnr∅ the map (Mop )op )⊛ → S × V⊗ of cocartesian fibrations over S classifies the functor
ζVrev ○ ζV ○ τ ∶ S → V BEnr∅ . Therefore the equivalence σ ′ classifies an equivalence id → ζVrev ○ ζV of
functors V BEnr∅ → V BEnr∅ .

Next we deduce naturality of the left V-enriched Yoneda-embedding ρM ∶ M⊛ → EnrFun∅,V (Mop , V)⊛
(Corollary 5.47):
Definition 5.46. Let V⊗ → Ass be a monoidal ∞-category compatible with small colimits. Let PV
be the composition
EnrFun∅,V (−,V)⊛
Ð→ ∅ ωBEnrV ÐÐÐÐÐÐÐÐÐÐ→ (V ωBEnrR
ζV L
op ̂ ⊂ V ω BEnr
̂ ∅.
V ωBEnr∅ ∅) ≃ V ω BEnr∅
Corollary 5.47. Let V⊗ → Ass be a monoidal ∞-category compatible with small colimits. There is a
natural transformation id → PV of functors V ωBEnr∅ → V ω BEnr
̂ ∅ whose component at any left en-
riched ∞-category M → V is the left V-enriched Yoneda-embedding ρM ∶ M⊛ → EnrFun∅,V (Mop , V)⊛ .
⊛ ⊗
Proof. By Construction 5.44 there is an embedding ζVrev ○ ζV ↪ PV of functors V ωBEnr∅ → V ω BEnr

̂ ∅.
By Proposition 5.45 there is a canonical equivalence id → ζVrev ○ ζV of endofunctors of V BEnr∅ .
We obtain an embedding id ↪ PV of functors V ωBEnr∅ → V ω BEnr ̂ ∅ whose component at any left
enriched ∞-category M → V is the left V-enriched Yoneda-embedding ρM ∶ M⊛ → ((Mop )op )⊛ ⊂
⊛ ⊗
EnrFun∅,V (Mop , V)⊛ by Proposition 5.45.

Remark 5.48. Corollary 5.47 gives naturality of the enriched Yoneda-embedding. Naturality of the
enriched Yoneda-embedding in Hinich’s model of enrichment is the topic of [1].
Next we apply Corollary 4.50 to deduce that our both models for the left enriched ∞-category of
enriched presheaves are equivalent:
Theorem 5.49. Let M⊛ → V⊗ be a left enriched ∞-category.
(1) The left V-enriched Yoneda-embedding
ρ ∶ M⊛ → EnrFun∅,V (Mop , V)⊛
uniquely extends to a left V-enriched equivalence
θ ∶ PLEnv(M)⊛ op ⊛
LEnr ≃ EnrFun∅,V (M , V) .
(2) Let V⊗ → Ass be a presentably monoidal ∞-category. The left V-enriched equivalence θ corre-
sponds to the V, V-enriched functor
LEnr × (M ) ⊂ PLEnv(M)LEnr × (PLEnv(M)LEnr ) ÐÐÐÐÐÐÐÐÐÐÐÐÐ→ V .

LMorPLEnv(M)LEnr (−,−)
op
PLEnv(M)⊛ op ⊛ ⊛ ⊛ ⊛
Proof. (1): By Corollary 4.50 it suffices to see that the essential image of ρ generates the target of ρ
under small colimits and left tensors and that for every Z ∈ M the left V-enriched functor
LMorEnrFun∅,V (Mop ,V) (ρ(Z), −) ∶ EnrFun∅,V (Mop , V)⊛ → V⊛
preserves small colimits and is V-linear. This holds by Corollary 5.24.
(2): The V, V-enriched functor
LEnr × (M ) ⊂ PLEnv(M)LEnr × (PLEnv(M)LEnr ) ÐÐÐÐÐÐÐÐÐÐÐÐÐ→ V

op
PLEnv(M)⊛ op ⊛ ⊛ ⊛ ⊛
corresponds to a left V-enriched functor θ′ ∶ PLEnv(M)⊛ LEnr → EnrFun∅,V (M , V) . By Proposition

op ⊛
′
4.46, Proposition 5.15 and Lemma 2.52 the left V-enriched functor θ preserves small colimits and is
′
left V-linear. Hence it is enough to see that θ∣M⊛ ≃ θ∣M⊛ by Proposition 2.109. Both latter enriched
functors correspond to the V, V-enriched functor
LMorM (−, −) ∶ M⊛ × (Mop )⊛ ⊂ PLEnv(M)⊛

LEnr × (PLEnv(M)LEnr ) ÐÐÐÐÐÐÐÐÐÐÐÐÐ→ V .
op ⊛ ⊛
Corollary 5.50. Let V⊗ → Ass be a monoidal ∞-category compatible with small colimits, M⊛ →
V⊗ , N⊛ → V⊗ small left enriched ∞-categories and F ∶ M⊛ → N⊛ a left V-enriched functor. There is a
commutative square of ∞-categories left tensored over V:
F̄
PLEnv(M)⊛
LEnr
/ PLEnv(N)⊛
LEnr
≃ ≃
(Fop )!

EnrFun∅,V (Mop , V)⊛ / EnrFun∅,V (Nop , V)⊛ .
5.4. Enriched ∞-categories of enriched functors as an internal hom. Next we prove that
the weakly enriched ∞-categories of enriched functors of Notation 5.9 are in fact pseudo-enriched,
enriched ∞-categories, respectively, under reasonable conditions (Theorem 5.51), and identify such as
the internal hom for the tensor product of enriched ∞-categories (Theorem 5.61).
Theorem 5.51. (1) Let V⊗ → Ass be a small ∞-operad, (W⊗ → Ass, T) a small localization pair,
M → V a small weakly left enriched ∞-category and N⊛ → V⊗ ×W⊗ a small right T-enriched
⊛ ⊗
∞-category. Then EnrFunV,∅ (M, N)⊛ → W⊗ is right T-enriched.

(2) Let V⊗ → Ass be an ∞-operad, M⊛ → V⊗ a small weakly left enriched ∞-category, N⊛ → V⊗ ×
W⊗ a right enriched ∞-category and W⊗ → Ass an ∞-operad that exhibits W as right enriched
in W such that W admits small limits and for every W ∈ W the functor RMorW (W, −) ∶ W → W
preserves small limits. Then EnrFunV,∅ (M, N)⊛ → W⊗ is a right enriched ∞-category.
Proof. (1): The V, W-enriched embedding N⊛ ↪ BEnv(N)
̃ T induces a right W-enriched embedding
EnrFunV,∅ (M, N)⊛ → EnrFunV,∅ (M, PBEnv(N)

̃ ⊛
T) .
Since N′⊛ ∶= V⊗ ×PEnv(V)⊗ PBEnv(N)⊛ T → V × T PEnv(W) is a presentably right tensored ∞-

⊗ −1 ⊗
category, by Corollary 5.14 and Lemma 2.52 also EnrFunV,∅ (M, N′ )⊛ → T−1 PEnv(W)⊗ is a pre-
sentably right tensored ∞-category and so a right enriched ∞-category. Thus by Proposition 4.30 the
pullback W⊗ ×T−1 PEnv(W)⊗ EnrFunV,∅ (M, N′ )⊛ ≃ EnrFunV,∅ (M, PBEnv(N)
̃ ⊛
T ) is a right T-enriched
∞-category, where the equivalence is by Remark 5.12. Thus also EnrFunV,∅ (M, N)⊛ → W⊗ is a right
T-enriched ∞-category.
(2): By Proposition 4.30 the weakly left enriched ∞-category M⊛ → V⊗ is the pullback of a unique
left enriched ∞-category M̄⊛ → PEnv(V)
̂ ⊗ ̂
. The following right PEnv(W)-enriched functor is an
equivalence since it is linear by Corollary 5.14 and induces an equivalence on underlying ∞-categories
by Proposition 4.30:
Θ⊛ ∶= EnrFun ̂ PEnv(V),∅
̂
(M̄, PBEnv(N)) ⊛
→ EnrFunV,∅ (M, PBEnv(N))
̂ ⊛
.
We obtain a right W-enriched embedding:
̂BEnv(N))
EnrFunV,∅ (M, N)⊛ ⊂ EnrFunV,∅ (M, P ̃ ⊛
≃ W⊗ ×PEnv(W)
̂
⊛
⊗ Θ .
By Corollary 5.14 and Lemma 2.52 the weakly right enriched ∞-category Θ⊛ → PEnv(W) ̂ ⊗
is a
presentably right tensored ∞-category and so exhibits Θ as right PEnv(W)-enriched. To prove (2)
̂
it is enough to see that for any G, H ∈ EnrFunV,∅ (M, N) the morphism object RMorΘ (ι ○ G, ι ○ H) ∈
̂
PEnv(W) belongs to W, where ι ∶ N⊛ → PBEnv(N)
̂ ⊛
is the canonical embedding.
̂
The following right PEnv(W)-enriched functor is an equivalence since it is linear by Corollary 5.14
and induces an equivalence on underlying ∞-categories by Corollary 4.36:
Θ⊛ = EnrFun ̂PEnv(V),∅
̂
(M̄, PBEnv(N)) ⊛
→ EnrFunPEnv(V),∅ (M̄, PBEnv(N))
̂ ⊛
.
̂
By Remark 5.19 the linear small colimits preserving embedding PBEnv(N)⊛ ⊂ PBEnv(N) ⊛
induces a
right PEnv(W)-linear small colimits preserving embedding
Θ′⊛ ∶= EnrFunPEnv(V),∅ (M̄, PBEnv(N))⊛ ⊂ EnrFunPEnv(V),∅ (M̄, PBEnv(N))
̂ ⊛
≃ PEnv(W)⊗ ×PEnv(W)
̂
⊛
⊗ Θ .
Let Λ ⊂ Θ be the full subcategory of all G ∈ Θ such that for every H ∈ EnrFunV,∅ (M, N) the
morphism object RMorΘ (G, ι ○ H) ∈ PEnv(W)̂ belongs to W. We will prove that Θ′ ⊂ Λ. We prove first
that Λ is closed in Θ under small colimits. This follows from the fact that the functor RMorΘ (−, −) ∶
Θop × Θ → PEnv(W)
̂ ̂
preserves limits component-wise and that W is closed in PEnv(W) under small
limits since W admits small limits and both functors W ⊂ Env(W), Env(W) ⊂ PEnv(W) ̂ preserve
arbitrary limits: the second functor does since it is the Yoneda-embedding, the first functor does
because for every W1 , ..., Wm for m ≥ 0 the functor MulW (W1 , ..., Wm , −) ∶ W → ̂ S preserves limits as
it factors as W(Wm , −) ○ RMorW (Wm−1 , −) ○ ... ○ RMorW (W1 , −) ∶ W → ̂ S. Thus Λ is closed in Θ under
small colimits.
̂
By Theorem 5.22 for every X ∈ M there is a PEnv(W)-enriched adjunction FX ∶ PBEnv(N)
̂ ⊛
⇄
Θ ∶ evX , where the right adjoint evaluates at X and the left adjoints sends Z ∈ PBEnv(N) to
⊛ ̂
LMorM (X, −) ⊗ Z. Since M is locally small and the enriched embedding PBEnv(N)⊛ ⊂ PBEnv(N) ̂ ⊛
is
linear, the latter adjunction restricts to a PEnv(W)-enriched adjunction PBEnv(N) ⇄ Θ , where the
⊛ ′⊛
right adjoint evaluates at X. Since M is small, by Corollary 5.23 the ∞-category Θ′ is generated under
small colimits by the objects FX (Z) for X ∈ M, Z ∈ PBEnv(N). Because PBEnv(N) is by definition
generated under small colimits by BEnv(N), the ∞-category Θ′ is generated under small colimits by
the objects FX (V1 ⊗ ... ⊗ Vn ⊗ Z ⊗ W1 ⊗ ... ⊗ Wm ) for X ∈ M, Z ∈ N, V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W
for n, m ≥ 0. Since the embedding Θ′ ⊂ Θ preserves small colimits, Θ′ is the smallest full subcategory
of Θ closed under small colimits containing the objects FX (V1 ⊗ ... ⊗ Vn ⊗ Z ⊗ W1 ⊗ ... ⊗ Wm ) for
X ∈ M, Z ∈ N, V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0. Consequently, Θ′ ⊂ Λ if these objects belong
to Λ, i.e. for every X ∈ M, Z ∈ N, V1 , ..., Vn ∈ V, W1 , ..., Wm ∈ W for n, m ≥ 0 the morphism object
RMorΘ (FX (V1 ⊗ ... ⊗ Vn ⊗ Z ⊗ W1 ⊗ ... ⊗ Wm ), ι ○ H) ≃
RMorPBEnv(N)
̂ (V1 ⊗ ... ⊗ Vn ⊗ Z ⊗ W1 ⊗ ... ⊗ Wm , H(X))
belongs to W. We first reduce to the case that m = 0. For that we note that the canonical embedding
j ∶ W⊛ → PEnv(W)
̂ ⊛
preserves right morphism objects since W is right enriched in itself: for every
W1 , ...Wm , X, Y ∈ W for m ≥ 0 the induced map
1 ) ⊗ ... ⊗ j(Wm , ), j(RMorW (X, Y))) →
̂
PEnv(W)(j(W
1 ) ⊗ ... ⊗ j(Wm ), RMorPEnv(W) (j(X), j(Y)))

̂
PEnv(W)(j(W
identifies with the canonical equivalence
MulPEnv(W)
̂ (j(W1 ), ..., j(Wm ); j(RMorW (X, Y))) ≃ MulW (W1 , ..., Wm ; RMorW (X, Y))
≃ MulW (W1 , ..., Wm , X; Y) ≃ MulPEnv(W)

̂ (j(W1 ), ..., j(Wm ), j(X); j(Y)).
The morphism object RMorPBEnv(N)
̂ (V1 ⊗ ... ⊗ Vn ⊗ Z ⊗ W1 ⊗ ... ⊗ Wm , H(X)) is the image of
RMorPBEnv(N)
̂ (V1 ⊗ ... ⊗ Vn ⊗ ι(Z), H(X))
under the functor
RMorPEnv(W)
̂ (j(Wm ), −) ○ ... ○ RMorPEnv(W)
̂ (j(W1 ), −) ∶ PEnv(W)
̂ → PEnv(W).
̂
Since j ∶ W⊛ → PEnv(W)
̂ ⊛
preserves right morphism objects, we can assume that m = 0. In this
case we like to see that RMorPBEnv(N)
̂ (V1 ⊗ ... ⊗ Vn ⊗ Z, H(X)) belongs to W. This follows from the
following equivalence:
RMorPBEnv(N)
̂ (V1 ⊗ ... ⊗ Vn ⊗ Z, H(X)) ≃ RMulMorPBEnv(N)
̂ (V1 , ..., Vn , Z, H(X))
≃ RMulMorN (V1 , ..., Vn , Z, H(X)).

Corollary 5.52. Let κ be a regular cardinal, V⊗ → Ass an ∞-operad, W⊗ → Ass a monoidal ∞-

category compatible with κ-small colimits, M⊛ → V⊗ a weakly left enriched ∞-category and N⊛ →
V⊗ × W⊗ a right κ-enriched ∞-category. Then EnrFunV,∅ (M, N)⊛ → W⊗ is right κ-enriched.
Corollary 5.53. Let V⊗ → Ass, W⊗ → Ass be small monoidal ∞-categories. The functor
V PLEnr∅ × ∅ PREnrW → V PBEnrW (M⊛ → V⊗ , N⊛ → W⊗ ) ↦ M⊛ × N⊛ → V⊗ × W⊗
admits component-wise right adjoints.
Construction 5.54. Let V⊗ → Ass, W⊗ → Ass be monoidal ∞-categories and µ ∶ V⊗ ×Ass W⊗ → W⊗
a monoidal functor. By Proposition 4.74 the pullback µ∗ (N)⊛ → V⊗ ×Ass W⊗ along µ is the pullback
of a unique bi-pseudo-enriched ∞-category Nµ⊛ → V⊗ × (Wrev )⊗ . Let
̃ ⊛ → W⊗ × (Vrev )⊗
N µ
be the corresponding bi-pseudo-enriched ∞-category via Notation 2.60.

Remark 5.55. If for every V ∈ V the functor µ(V, −) ∶ W → W admits a right adjoint ΓV and
N⊛ → W⊗ exhibits N as right enriched in W, then Nµ⊛ → V⊗ × (Wrev )⊗ exhibits N as right enriched in
Wrev , where the right multi-morphism object for V1 , ..., Vn ∈ V, X, Y ∈ N is
RMulMorNµ (V1 , ..., Vn , X, Y) = ΓV1 ⊗...⊗Vn (RMorN (X, Y)).
If for every W ∈ W the functor µ(−, W) ∶ V → W admits a right adjoint ΞW and N⊛ → W⊗ exhibits
N as right enriched in W, then Nµ⊛ → V⊗ × (Wrev )⊗ exhibits N as left enriched in V, where the left
multi-morphism object for W1 , ..., Wm ∈ W, X, Y ∈ N is
LMulMorNµ (X, W1 , ..., Wm , Y) = ΞW1 ⊗...⊗Wm (RMorN (X, Y)).
Notation 5.56. Let V⊗ → Ass, W⊗ → Ass be monoidal ∞-categories and µ ∶ V⊗ ×Ass W⊗ → W⊗ a
monoidal functor. Let N⊛ → W⊗ be a left pseudo-enriched ∞-category and M⊛ → V⊗ , O⊛ → W⊗
weakly left enriched ∞-categories.
(1) Let
EnrFunV (M, N)⊛ → W⊗
be the weakly left enriched ∞-category corresponding to the weakly right enriched ∞-category
EnrFunV,∅ (M, Nµ )⊛ → (Wrev )⊗ .
(2) Let
EnrFunW (O, N)⊛ → V⊗
be the weakly left enriched ∞-category corresponding to the weakly right enriched ∞-category
̃ µ )⊛ → (Vrev )⊗ .
EnrFunW,∅ (O, N
Theorem 5.51 has the following consequences:
Corollary 5.57. Let κ be a small regular cardinal, V⊗ → Ass, W⊗ → Ass monoidal ∞-categories
compatible with κ-small colimits and µ ∶ V⊗ ×Ass W⊗ → W⊗ a monoidal functor preserving κ-small
colimits component-wise. Let N⊛ → W⊗ be a left κ-enriched ∞-category and M⊛ → V⊗ , O⊛ → W⊗
weakly left enriched ∞-categories. The following weakly left enriched ∞-categories are left κ-enriched:
EnrFunV (M, N)⊛ → W⊗ , EnrFunW (O, N)⊛ → V⊗ .
Corollary 5.58. Let V⊗ → Ass, W⊗ → Ass be monoidal ∞-categories, µ ∶ V⊗ ×Ass W⊗ → W⊗ a
monoidal functor, M⊛ → V⊗ , O⊛ → W⊗ small weakly left enriched ∞-categories and N⊛ → W⊗ a left
(1) If W admits small limits, the monoidal structure on W is closed and for every V ∈ V the
functor µ(V, −) ∶ W → W admits a right adjoint, the weakly left enriched ∞-category
EnrFunV (M, N)⊛ → W⊗
is a left enriched ∞-category.
(2) If V admits small limits, the monoidal structure on V is closed and for every W ∈ W the
functor µ(−, W) ∶ V → W admits a right adjoint, the weakly left enriched ∞-category
EnrFunW (O, N)⊛ → V⊗
is a left enriched ∞-category.
Proposition 5.59. Let V⊗ → Ass, W⊗ → Ass be monoidal ∞-categories and µ ∶ V⊗ ×Ass W⊗ → W⊗
a monoidal functor. Let N⊛ → W⊗ be a left pseudo-enriched ∞-category and M⊛ → V⊗ , O⊛ → W⊗
weakly left enriched ∞-categories. There are canonical equivalences
EnrFunW,∅ (O, EnrFunV (M, N)) ≃ EnrFunW,∅ (µ! (M × O), N) ≃ EnrFunV,∅ (M, EnrFunW (O, N)).
Proof. There is a canonical equivalence
EnrFunW,∅ (µ! (M × O), N) ≃ EnrFunV×W,∅ (M × O, µ∗ (N)) ≃ EnrFunV,Wrev (M × Orev , Nµ ).
There are canonical equivalences:
EnrFunW,∅ (O, EnrFunV (M, N)) ≃ EnrFun∅,Wrev (Orev , EnrFunV,∅ (M, Nµ )) ≃
EnrFunV,Wrev (M × Orev , Nµ ),
EnrFunV,∅ (M, EnrFunW (O, N)) ≃ EnrFun∅,Vrev (Mrev , EnrFunW,∅ (O, N ̃ µ )) ≃
̃ µ ) ≃ EnrFunV,Wrev (M × Orev , Nµ ).
EnrFunW,Vrev (O × Mrev , N

Notation 5.60. Let V → Ass be a monoid in the ∞-category Mon corresponding to a braided
⊗
monoidal ∞-category and W⊗ → Ass a monoidal ∞-category equipped with a left action of V⊗ → Ass
in Mon. Let M⊛ → V⊗ , N⊛ → V⊗ , O⊛ → W⊗ be small left enriched ∞-categories.
(1) Let (M ⊗ N)⊛ → V⊗ be the tranfer of enrichment of the left V × V-enriched ∞-category
M⊛ ×Ass N⊛ → V⊗ ×Ass V⊗ along the monoidal tensor product functor V⊗ ×Ass V⊗ → V⊗ .
(2) Let (M ⊗ O)⊛ → W⊗ be the tranfer of enrichment of the left V × W-enriched ∞-category
M⊛ ×Ass N⊛ → V⊗ ×Ass W⊗ along the left action functor V⊗ ×Ass W⊗ → W⊗ .
By [3], [8, Corollary 5.7.12.] for every braided monoidal ∞-category V and monoidal ∞-category
W⊗ → Ass with a left V-action the ∞-category V LEnr carries a monoidal structure and W LEnr is left
tensored over V LEnr. The tensor product and left action are defined as in Notation 5.60.
Theorem 5.61. Let V⊗ → Ass be a closed braided monoidal ∞-category and W⊗ → Ass a closed
monoidal ∞-category equipped with a left action of V⊗ → Ass such that V, W admit small limits and
the tensor product V × V → V and left action V × W → W admit component-wise right adjoints.
(1) Let M⊛ → V⊗ , N⊛ → V⊗ , O⊛ → V⊗ be small left enriched ∞-categories. The small weakly
left enriched ∞-categories EnrFunV (M, N)⊛ → V⊗ , EnrFunV (O, N)⊛ → V⊗ are left enriched
∞-categories and there are canonical equivalences
EnrFunV,∅ (O, EnrFunV (M, N)) ≃ EnrFunW,∅ (M ⊗ O, N) ≃ EnrFunV,∅ (M, EnrFunV (O, N)).
(2) Let M⊛ → V⊗ , N⊛ → V⊗ , O⊛ → W⊗ be small left enriched ∞-categories. The small weakly
left enriched ∞-categories EnrFunV (M, N)⊛ → W⊗ , EnrFunW (O, N)⊛ → V⊗ are left enriched
∞-categories and there are canonical equivalences
EnrFunW,∅ (O, EnrFunV (M, N)) ≃ EnrFunW,∅ (M ⊗ O, N) ≃ EnrFunV,∅ (M, EnrFunW (O, N)).
Every monoidal ∞-category V⊗ → Ass compatible with small colimits carries a canonical left action
in ccMon of the ∞-category S of small spaces equipped with the cartesian structure. The left action
functor is the canonical monoidal functor µ ∶ S × V → S ⊗ V ≃ V.
Corollary 5.62. Let V⊗ → Ass be a presentably monoidal ∞-category. Let M⊛ → V⊗ , N⊛ → V⊗ be
left V-enriched ∞-categories and K ∈ Cat∞ . There is a canonical equivalence
Fun(K, EnrFunV,∅ (M, N)) ≃ EnrFunV (µ! (K × M), N).
Proof. By Corollary 4.33 (3) the forgetful functor S LEnr∅ → Cat∞ is an equivalence. Let K⊛ → S× be
the unique left enriched ∞-category whose underlying ∞-category is K.
By Proposition 5.59 and Corollary 4.33 there is a canonical equivalence
EnrFunV,∅ (µ! (K × M), N) ≃ EnrFunS,∅ (K, EnrFunV (M, N)) ≃ Fun(K, EnrFunV,∅ (M, N)).

5.5. Enriched profunctors.

Definition 5.63. Let V⊗ → Ass be a monoidal ∞-category compatible with small colimits and
C⊛ → V⊗ , D⊛ → V⊗ right V-enriched ∞-categories. A V-enriched profunctor C → D is a V, V-enriched
functor (Dop )⊛ × C⊛ → V⊛ .
Remark 5.64. Haugseng [6] defines enriched bimodules, another model for enriched profunctors.
Hinich [14, 8.1.] defines enriched profunctors under the name enriched correspondences and identifies
such with enriched functors to the interval [14, Proposition 8.3.2.].
Remark 5.65. Let C⊛ → V⊗ , D⊛ → V⊗ be right enriched ∞-categories. A V-enriched profunctor
C → D corresponds to a right V-enriched functor C⊛ → EnrFunV,∅ (Dop , V)⊛ , which corresponds
to a right V-linear small colimits preserving functor EnrFunV,∅ (Cop , V)⊛ → EnrFunV,∅ (Dop , V)⊛ by
Theorem 5.49.
Example 5.66. Let C⊛ → V⊗ be a right enriched ∞-category. The V, V-enriched morphism ob-
ject functor LMor ∶ (Cop )⊛ × C⊛ → V⊛ of Notation 5.36 is a V-enriched profunctor C → C, which
corresponds to the identity of EnrFunV,∅ (Cop , V)⊛ and to the right V-enriched Yoneda-embedding
C⊛ → EnrFunV,∅ (Cop , V)⊛ .
Example 5.67. Let M⊛ → V⊗ be a right tensored ∞-category compatible with small colimits and
A, B associative algebras in V. By [16, Theorem 4.8.4.1.] and Corollary 4.52 there is a canonical left
V-enriched equivalence
EnrFun∅,V (BA, M)⊛ ≃ LinFunL∅,V (LModA (V), M)⊛ ≃ RModA (M)⊛ .
Hence there is a canonical left V-enriched equivalence EnrFun∅,V (BA, V)⊛ ≃ RModA (V)⊛ and so
dually a right V-enriched equivalence EnrFunV,∅ (BAop , V)⊛ ≃ LModA (V)⊛ .
Thus a V-enriched profunctor BA → BB corresponds to a right V-linear small colimits preserving
functor LModA (V)⊛ → LModB (V)⊛ , which corresponds to an object of RModA (LModB (V)) identified
with an A, B-bimodule in V by Remark 4.83.
Definition 5.68. Let C⊛ → V⊗ , D⊛ → V⊗ , E⊛ → V⊗ be right enriched ∞-categories and F ∶ C → D, G ∶
D → E be V-enriched profunctors corresponding to right V-linear small colimits preserving functors
PV (C)⊛ → PV (D)⊛ , PV (D)⊛ → PV (E)⊛ . The relative tensor product of F and G, denoted by F⊗D G, is
the V-enriched profunctor C → E corresponding to the right V-linear small colimits preserving functor
PV (C)⊛ → PV (D)⊛ → PV (E)⊛ .
Example 5.69. There are canonical equivalences F ⊗D LMorD ≃ F, LMorD ⊗D G ≃ G.
Proposition 5.70. Let C⊛ → V⊗ , D⊛ → V⊗ , E⊛ → V⊗ be right enriched ∞-categories and F ∶ C →
D, G ∶ D → E be V-enriched profunctors. For every X ∈ C, Y ∈ E there is a canonical equivalence
(F⊗D G)(Y, X) ≃ colim (colimZ1 ,....,Zn ∈D≃ F(Z1 , X)⊗LMorD (Z2 , Z1 )⊗....⊗LMorD (Zn , Zn−1 )⊗G(Y, Zn ))).
[n]∈∆op
Proof. By Proposition 5.26 there is a canonical equivalence

F(−, X) ≃ colim
op
(colimZ1 ,....,Zn ∈D≃ F(X, Z1 ) ⊗ LMorD (Z2 , Z1 ) ⊗ .... ⊗ LMorD (Zn , Zn−1 ) ⊗ LMorD (−, Zn ))
[n]∈∆
in EnrFunV,∅ (Dop , V) and so a canonical equivalence in EnrFunV,∅ (Eop , V) ∶

(F ⊗D G)(−, X) ≃ Ḡ(F(, −X)) ≃
colim (colimZ1 ,....,Zn ∈D≃ F(Z1 , X) ⊗ LMorD (Z2 , Z1 ) ⊗ .... ⊗ LMorD (Zn , Zn−1 ) ⊗ Ḡ(LMorD (−, Zn )))
[n]∈∆op
≃ colim
op
(colimZ1 ,....,Zn ∈D≃ F(Z1 , X) ⊗ LMorD (Z2 , Z1 ) ⊗ .... ⊗ LMorD (Zn , Zn−1 ) ⊗ G(−, Zn ))).
[n]∈∆

5.6. An end formula for morphism objects of enriched functor ∞-categories. In the following
we describe morphism objects in the enriched ∞-category of enriched functors as an enriched end
(Theorem 5.78).
In the following we will use symmetric monoidal ∞-categories [16, Definition 2.0.0.7.], which are
cocartesian fibrations V⊠ → Comm to the category Comm of finite pointed sets. There is a canonical
functor θ ∶ Ass → Comm, [n] ↦ {1, ..., n, ∗} [16, Construction 4.1.2.9.] and we write V⊗ → Ass for the
pullback of a symmetric monoidal ∞-category V⊠ → Comm along θ, which is a monoidal ∞-category.
If V⊗ → Ass is the pullback of a symmetric monoidal ∞-category V⊠ → Comm, there is a canonical
monoidal equivalence (Vrev )⊗ ≃ V⊗ since θ factors as Ass ÐÐÐ→ Ass Ð → Comm. Consequently, there
(−)op θ
is no need to distinguish between (weakly) left and right V-(pseudo)-enriched ∞-categories, which we
therefore call (weakly) V-(pseudo)-enriched ∞-categories. Moreover there is no need to distinguish
between left and right morphism objects, which we call morphism objects, and between left and right
(co)tensors, which we call (co)tensors.
Notation 5.71. Let V⊠ → Comm be a symmetric monoidal ∞-category and M⊛ → V⊗ a right V-
enriched ∞-category. Let
MorM ∶ (Mop ⊗ M)⊛ → V⊛
be the left V-enriched functor corresponding to the left V × V-enriched functor (Mop )⊛ ×Ass M⊛ →
⊗∗V (V)⊛ corresponding to the V, V-enriched functor RMorM ∶ (Mop )⊛ × M⊛ → V⊛ of Notation 5.36.
Definition 5.72. Let V⊗ → Comm be a symmetric monoidal ∞-category, M⊛ → V⊗ , J⊛ → V⊗ pseudo-
enriched ∞-categories and F ∶ (Jop ⊗ J)⊛ → M⊛ a V-enriched functor.
● The V-enriched end of F, denoted by ∫J F, is the object representing the V-enriched presheaf
Mop → V, X ↦ MorEnrFunV (Jop ⊗J,V) (MorJ , (MorM (X, −)) ○ F).
● The V-enriched coend of F, denoted by ∫ F, is the object corepresenting the V-enriched

J
functor
M → V, X ↦ MorEnrFunV (J⊗Jop ,V) (MorJop , MorM (−, X) ○ Fop ).
Remark 5.73. By universal property the V-enriched end of F is the V-enriched coend of Fop .
Remark 5.74. Let V⊗ → Comm be a symmetric monoidal ∞-category, J⊛ → V⊗ a pseudo-enriched
∞-category and F ∶ (Jop ⊗ J)⊛ → V⊛ a V-enriched functor. By Definition 5.72 there is an equivalence
∫ F ≃ MorEnrFunV (Jop ⊗J,V) (MorJ , F).

J
Remark 5.75. Let V → Comm be a symmetric monoidal ∞-category, M⊛ → V⊗ , J⊛ → V⊗ pseudo-

⊗
enriched ∞-categories and F ∶ (Jop ⊗ J)⊛ → M⊛ a V-enriched functor. A V-enriched functor φ ∶ M⊛ →

N⊛ that admits a V-enriched right adjoint γ preserves V-enriched coends: the canonical morphism
∫ φ ○ F → φ(∫J F) is corepresented by the canonical equivalence
J
MorEnrFunV (J⊗Jop ,V) (MorJop , MorM (−, γ(X))○Fop ) → MorEnrFunV (J⊗Jop ,V) (MorJop , MorN (−, X)○φop ○Fop ).
Notation 5.76. Let J be an ∞-category. Let q ∶ Tw(J) → Jop × J be the left fibration classifying the
mapping space functor J(−, −) ∶ Jop × J → S.
Lemma 5.77. Let M⊛ → S× be an enriched ∞-category. The S-enriched end of a S-enriched functor
F ∶ (Jop )⊛ ×Ass J⊛ ≃ (Jop ⊗ J)⊛ → M⊛ is the limit of the functor
Tw(J) Ð
→ Jop × J Ð
→ M.
q F
Proof. For every X ∈ M let X → Jop × J be the left fibration classifying the functor M(X, −) ○ F ∶
Jop × J → S. By Remark 5.74 there is a canonical equivalence
M(X, ∫ F) ≃ Fun(Jop × J, S)(J(−, −), M(X, F(−))) ≃ FunJop ×J (Tw(J), X) ≃

J
FunTw(J) (Tw(J), Tw(J) ×Jop ×J X) ≃ lim(M(X, −) ○ F ○ q) ≃ M(X, lim(F ○ q))

representing an equivalence ∫J F ≃ lim(F ○ q).

Theorem 5.78. Let V⊗ → Comm be a presentably symmetric monoidal ∞-category, M⊛ → V⊗ a

small enriched ∞-category and J⊛ → V⊗ a small weakly enriched ∞-category. Let F, G ∶ J⊛ → M⊛ be
V-enriched functors. Then MorEnrFunV (J,M) (F, G) is the V-enriched end ∫J MorM ○ (Fop ⊗ G).
Proof. We use Remark 5.74. There is a chain of equivalences:

MorEnrFunV (Jop ⊗J,V) (MorJ , MorM ○ (Fop ⊗ G)) ≃
MorEnrFunV (J,EnrFunV (Jop ,V)) (ρJ , EnrFunV (Fop , V) ○ ρM ○ G) ≃

MorEnrFunV (J,EnrFunV (Mop ,V)) (F! ○ ρJ , ρM ○ G) ≃
MorEnrFunV (J,EnrFunV (Mop ,V)) (ρM ○ F, ρM ○ G) ≃ MorEnrFunV (J,M) (F, G).
The first equivalence is by definition of the V-enriched functor EnrFunV (Fop , V) ∶ EnrFunV (Mop , V)⊛ →
EnrFunV (Jop , V)⊛ of Notation 5.17 and Notation 5.56. The second equivalence is by adjointness. The
third equivalence is by Corollary 5.50. The fourth equivalence holds by Lemma 5.7 and the fact that
the enriched Yoneda-embedding is an embedding by Proposition 5.40.

Corollary 5.79. Let V⊗ → Comm be a presentably symmetric monoidal ∞-category, J⊛ → V⊗ a

small enriched ∞-category and M⊛ → V⊗ an enriched ∞-category that admits small conical colimits
and tensors. Every V-enriched functor F ∶ J⊛ → M⊛ is the V-enriched coend of the V-enriched functor
⊗ ○ (F ⊗ ιJop ) ∶ (J ⊗ Jop )⊛ ÐÐÐÐ→ (M ⊗ EnrFunV (J, V))⊛ Ð

→ EnrFunV (J, M)⊛ ,
F⊗ιJop ⊗
where ⊗ is the V-enriched functor

(M ⊗ EnrFunV (J, V))⊛ ≃ (LinFunLV (V, M) ⊗ EnrFunV (J, V))⊛ Ð
→ EnrFunV (J, M)⊛ .
○
Proof. Let G ∶ J⊛ → M⊛ be a V-enriched functor. By the enriched Yoneda-lemma (Corollary 5.23) the
composition
(Jop ⊗ J)⊛ ÐÐÐÐÐÐÐÐÐ→ (EnrFunV (J, M)op )⊛ ÐÐÐÐÐÐÐÐÐÐÐÐÐ→ V⊛

(⊗○(F⊗ιJop ))op MorEnrFunV (J,M) (−,G)
is the V-enriched functor MorM ○ (Fop ⊗ G). Hence by Theorem 5.78 there is a canonical equivalence
MorEnrFunV (J,M) (F, G) ≃ MorEnrFunV (Jop ⊗J,V) (MorJ , MorM ○ (Fop ⊗ G)) ≃
MorEnrFunV (Jop ⊗J,V) (MorJ , MorEnrFunV (J,M) (−, G) ○ (⊗ ○ (F ⊗ ιJop ))op ).

Corollary 5.80. Let V⊗ → Comm be a presentably symmetric monoidal ∞-category, J⊛ → V⊗ a small

V-enriched ∞-category, N⊛ → V⊗ a V-enriched ∞-category that admits small conical colimits and
tensors and let F ∶ (Jop )⊛ → V⊛ , G ∶ J⊛ → N⊛ be V-enriched functors. Let Ḡ ∶ EnrFunV (Jop , V)⊛ → N⊛
be the unique V-enriched left adjoint extension Ḡ ∶ EnrFunV (Jop , V)⊛ → N⊛ of G. There is a canonical
equivalence in N ∶
J
Ḡ(F) ≃ ∫ ⊗ ○ (F ⊗ G).
Proof. By Corollary 5.79 there is a canonical equivalence F ≃ ∫ ⊗ ○ (F ⊗ ιJ ). Since Ḡ is a V-enriched

J
left adjoint, it preserves V-enriched coends. Moreover because Ḡ preserves tensors, the canonical map
⊗ ○ (F ⊗ (Ḡ ○ ιJ )) → Ḡ ○ (⊗ ○ (F ⊗ ιJ )) is an equivalence. We obtain a canonical equivalence
J J
Ḡ(F) ≃ Ḡ(∫ ⊗ ○ (F ⊗ ιJ )) ≃ ∫ Ḡ ○ (⊗ ○ (F ⊗ ιJ ))
J J
≃∫ ⊗ ○ (F ⊗ (Ḡ ○ ιJ )) ≃ ∫ ⊗ ○ (F ⊗ G).

Corollary 5.81. Let V⊗ → Comm be a presentably symmetric monoidal ∞-category, C⊛ → V⊗ , D⊛ →

V⊗ , E⊛ → V⊗ small V-enriched ∞-categories and F a V-enriched profunctor C → D and G a V-enriched
profunctor D → E. For every X ∈ C, Y ∈ E there is a canonical equivalence in V ∶
J
(F ⊗D G)(X, Y) ≃ ∫ ⊗ ○ (F(−, X) ⊗ G(Y, −).
References
[1] Shay Ben-Moshe. Naturality of the ∞-categorical enriched yoneda embedding. Journal of Pure and Applied Algebra,
228(6):107625, 2024.
[2] Timothy Campion. The gray tensor product of (∞, n)-categories. arxiv preprint arXiv: 2311.00205, 2023.
[3] David Gepner and Rune Haugseng. Enriched ∞-categories via non-symmetric ∞-operads. Advances in Mathemat-
ics, 279:575 – 716, 2015.
[4] David Gepner, Rune Haugseng, and Thomas Nikolaus. Lax colimits and free fibrations in ∞-categories. Documenta
Mathematica, 22, 01 2015.
[5] Rune Haugseng. Rectification of enriched ∞-categories. Algebr. Geom. Topol., 15(4):1931–1982, 2015.
[6] Rune Haugseng. Bimodules and natural transformations for enriched ∞-categories. Homology, Homotopy and
Applications, 18(1):71–98, 2016.
[7] Rune Haugseng. ∞-operads via symmetric sequences. Mathematische Zeitschrift, 301, 05 2022.
[8] Rune Haugseng. On the tensor product of enriched ∞-categories. arxiv preprint arXiv: 2311.13362, 2023.
[9] Hadrian Heine. Restricted L∞ -algebras. https://osnadocs.ub.uni-osnabrueck.de/handle/urn:nbn:de:gbv:700-
201909201996.
[10] Hadrian Heine. An equivalence between enriched ∞-categories and ∞-categories with weak action. Advances in
Mathematics, 417:108941, 2023.
[11] Hadrian Heine. A monadicity theorem for higher algebraic structures. arxiv preprint arXiv: 1712.00555, 2023.
[12] Hadrian Heine. A topological model for cellular motivic spectra. arxiv preprint arXiv: 1712.00521, 2023.
[13] Hadrian Heine. An equivalence between two models of ∞-categories of enriched presheaves. arxiv preprint arXiv:
2306.08115, 2024.
[14] Vladimir Hinich. Yoneda lemma for enriched ∞-categories. Advances in Mathematics, 367:107129, 2020.
[15] Vladimir Hinich. Colimits in enriched ∞-categories and day convolution. arxiv preprint arXiv: 2101.09538, 2021.
[16] Jacob Lurie. Higher Algebra. available at http://www.math.harvard.edu/ lurie/.
[17] Jacob Lurie. Higher topos theory, volume 170 of Annals of Mathematics Studies. Princeton University Press,
Princeton, NJ, 2009.
[18] Andrew W. Macpherson. The operad that co-represents enrichment. Homology Homotopy Appl., 23(1):387–401,
2021.

2406 09832v1

Uploaded by

Copyright:

Available Formats

2406 09832v1

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2406 09832v1

Uploaded by

Copyright:

Available Formats

ON BI-ENRICHED ∞-CATEGORIES

5.5. Enriched profunctors 87

(1) MulO (V1 ⊗ ... ⊗ Vn , X; Y) ≃ MulO (V1 , ..., Vn , X; Y).

In case that V is a presentably symmetric monoidal ∞-category, we prove a refinement of Corollary

● active if it corresponds to a map of ∆, which preserves the minimum and maximum.

Notation 2.4. For every ∞-operad φ ∶ V⊗ → Ass we call V ∶= V⊗ [1]

morphism [1] → [n] in ∆, where V ∈ V⊗

(1) We call φ locally small if the ∞-categories V⊗ , W⊗ , M⊛ are locally small.

Definition 2.44. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category.

Notation 2.48. Let M⊛ → V⊗ × W⊗ , N⊛ → V′⊗ × W′⊗ be weakly bi-enriched ∞-categories.

Notation 2.53. Let

Notation 2.58. Let

2.4. Enriched adjunctions. Next we define enriched adjunctions.

The morphism η ∶ id → G ○ F exhibits F as an enriched left adjoint of G (or G as an enriched right

Remark 2.73. Let O⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category, F ∶ M⊛ → N⊛ , G ∶ N⊛ → M⊛

(2) A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ right trivial if for every n ≥ 0, m > 0

V,W ∶= trivV × trivW → V × W .

Remark 2.76 gives the following example:

V,W ∶= trivV × K × trivW → V × W .

Remark 2.83. The weakly bi-enriched ∞-category K⊛

Proposition 2.84. Let N⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category.

EnrFunV,W (M × trivW , Ntriv ) ≃ EnrFunV,W (M × trivW , N).

Consequently, it is enough to see that the canonical functor

EnrFunV,W (M × trivW , Ntriv ) → EnrFunV,∅ (M, N)

is an equivalence. The functor N⊛ → V⊗ × W⊗ over W⊗ is a map of cocartesian fibrations relative to

So it suffices to show that the canonical functor

EnrFunV,W (M × trivW , N × trivW ) → EnrFunV,∅ (M, N)

W , N × trivW ) ≃ FunV⊗ ×triv⊛ (M × trivW , N × trivW )

≃ FunV⊗ (M⊛ × triv⊛ ⊛ ⊛ ⊛ ⊛

and the functor

evaluating at the final object of triv⊛

EnrFunV,W (M × trivW , N × trivW ) ≃ Fun′ (triv⊛

Corollary 2.85. Let V⊗ → Ass, W⊗ → Ass be small ∞-operads.

Definition 2.92. Let M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category.

AlgLIndκ (V) (W) → AlgκV (W).

restricts to an adjunction φ ∶ AlgV (Ind

Proof. We can assume that M⊛ → V⊗ × W⊗ , N⊛ → V⊗ × W⊗ are totally small. Let K be a small

EnrFunV,W (M, PBEnv(M)) corresponds to a V, W-enriched functor Kop × M⊛ → (PBEnv(M)⊛ )K ,

λ ∶ Cat∞ (Kop , EnrFunLPEnv(V),PEnv(W) (PBEnv(M), PBEnv(N))) ≃

Cat∞ (K, EnrFunR

representing the desired equivalence EnrFunLV,W (M, N)op ≃ EnrFunR

3. Enrichment and pseudo-enrichment

Definition 3.1. Let φ ∶ M⊛ → V⊗ × W⊗ be a weakly bi-enriched ∞-category.

Notation 3.4. Let

PEnv(V),PEnv(W) (PBEnv(M), PEnv(V) ⊗ PEnv(W))

≃ MulM (V1 , ..., Vn , X, W1 , ..., Wm ; Y).

Remark 3.11. Let M⊛ → V⊗ ×W⊗ be a weakly bi-enriched ∞-category and X, Y ∈ M, W1 , ..., Wm ∈ W

Notation 3.15. Let

Remark 3.19. Let C, D be small ∞-categories. A presheaf α ∈ P(C × D) is representable in both

Consequently, the canonical equivalence P(C × D) ≃ Fun(Cop , P(D)) restricts to an equivalence

LMorM (X, Y)op

Thus for every V ∈ V, W ∈ W there is a canonical equivalence

4. Enriched presheaves via tensored envelopes

if it is a monoidal functor preserving κ-small colimits.

(2) A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ exhibits M as right T-enriched if for every

S,T ⊂ S PEnv(V) ×PEnv(V)⊗ PBEnv(M) ×PEnv(W)⊗ T PEnv(W)

If S = ∅ or T = ∅, we drop S, T from the notation, respectively.

Remark 4.17. A weakly bi-enriched ∞-category M⊛ → V⊗ × W⊗ is a S, T-bi-enriched ∞-category if

EnrFunS−1 PEnv(V),T−1 PEnv(W)(PBEnv(M)S,T , N) → EnrFunV,W (M, N)