Homepage Redesign

The main AnythingLLM homepage has been completely redesigned to be more modern and user-friendly so you can instantly start chatting the second you open the app after onboarding.

Native Tool Calling

this only applies to local LLM providers. It has no impact on cloud LLMs like OpenAI, Anthropic, or Azure.

We have completely overhauled how @agent tool calling works. Now, we will leverage the new native tool calling abilities of your LLM provider and model.

What this means for you:

• You can now run complex, multi-step tool calls with your LLM provider and model.
• Your model will now continue to work until your final response is generated or determined to be complete.
• You will get 100x better responses from even small tool-calling models

We have implemented safeguards as well to prevent infinite loops with a maximum of 10 tool calls per response to prevent runaway tasks.

Limitations

Most providers do not allow us to probe for if a model supports native tool calling.

The following local LLM providers will automatically support native tool calling if your model supports it:

• Default Built in LLM Provider (AnythingLLM Default)
• Ollama
• LM Studio

For others, you will need to set an ENV variable to enable native tool calling for supported providers.

• Generic OpenAI
• Groq
• AWS Bedrock
• Lemonade
• LiteLLM
• Local AI
• OpenRouter

This can be set via the PROVIDER_SUPPORTS_NATIVE_TOOL_CALLING environment variable.

PROVIDER_SUPPORTS_NATIVE_TOOL_CALLING="bedrock,generic-openai,groq,lemonade,litellm,local-ai,openrouter"

Lemonade by AMD Integration

Lemonade by AMD is an open-source local model runtime that optimizes performance and efficiency for local models (LLM, ASR, TTS, Image Generation, etc.) for all types of hardware including AMD GPUs and NPUs.

We have added first class support so you can use your local models running via Lemonade within AnythingLLM for the best application experience on top of your local hardware.

What's Changed

• fix: typo in contribution guidelines, update project metadata and pull_request_temp...md by @dipanshurdev in #5010
• feat: update light mode UI sidebar by @angelplusultra in #4996
• fix(frontend): fix event listener memory leak in useIsDisabled hook by @dipanshurdev in #5027
• feat: dedicated dark theme option with system preference support by @angelplusultra in #5007
• Implement new home page redesign by @shatfield4 in #4931
• fix: GitLab connector infinite loop and rate limit crash for large repos by @angelplusultra in #5021
• fix: add password character validation to onboarding single-user setup by @angelplusultra in #5037
• Native Tool calling by @timothycarambat in #5071
• fix: resolve Gemini agent 400 error on tool call responses by @angelplusultra in #5054
• fix: prevent CMD/CTRL+Arrow scroll from overriding textarea cursor movement by @angelplusultra in #5053
• Normalize scraper runtimeargs for bulk-scraper by @timothycarambat in #5083
• Lemonade integration by @timothycarambat in #5077

New Contributors

• @dipanshurdev made their first contribution in #5010

Full Changelog: v1.11.0...v1.11.1

AnythingLLM Desktop overlay is live!

this is a free & desktop specific feature!

Now, AnythingLLM Desktop has an OS-level and application aware panel that opens in a single keystroke. Seamlessly ingest your current open applications alongside all other chat functionality you use like document chat, RAG, agents, and more.

This panel is such a smoother and more convenient way to use AnythingLLM - we highly recommend this for daily use!

What's Changed

• Update common.js by @j0rDy in #4894
• Add keyboard shortcuts to scroll to top and bottom of chat history by @angelplusultra in #4870
• 4825 patch yt file collector api by @timothycarambat in #4904
• Refine and Standardize Username Constraints by @angelplusultra in #4828
• New login page UI by @shatfield4 in #4862
• [Chore] Autotranslation tool using DMR by @timothycarambat in #4907
• Persist Ollama context preferences in LC tools by @timothycarambat in #4908
• Refactor Ollama context window setting by @timothycarambat in #4909
• Patch AzureOpenAI tool calling from function to tool by @timothycarambat in #4922
• Thinking block persist toggle state by @timothycarambat in #4916
• Manage Onboarding decision via DB flag by @timothycarambat in #4926
• Fix sidebar thread layer by @17ColinMiPerry in #4881
• Toggle switch component by @shatfield4 in #4890
• Fix: Clean username already exists error by @angelplusultra in #4914
• fix: prevent Citations UI glitching during streaming chats by @naaa760 in #4897
• Support PrivateModeAI Integration by @timothycarambat in #4937
• feat: Add ability to edit existing SQL agent connections by @angelplusultra in #4848
• Fix double scroll container in agent flow builder by @shatfield4 in #4933
• feat: Enable essential ESLint rules and refactor frontend lint config by @angelplusultra in #4923
• SambaNova Integration by @timothycarambat in #4943
• Update LMStudio LLM & Embedder for API token by @timothycarambat in #4948
• Web push notifications by @timothycarambat in #4942
• Fix agent workspace settings loading forever by @shatfield4 in #4956
• add provider field to chats by @timothycarambat in #4959
• UAE region support for bedrock models by @vigorous-ride in #4964
• fix: scroll active sidebar items into view by @angelplusultra in #4965
• refactor: Migrate system preferences to field-based endpoint and remove deprecated endpoint by @angelplusultra in #4958
• Refactor: Remove redundant message state sync in ChatContainer by @shatfield4 in #4941
• chore: bump ollama SDK from 0.5.10 to 0.6.3 by @angelplusultra in #4977
• set embedder output dimensions for LocalAI and Gemini (gemini-embedding-001) by @timothycarambat in #4980
• Fix Sidebar Clipping When Closed by @angelplusultra in #4970
• fix: correct TPS calculation for Generic OpenAI provider with llama.cpp by @angelplusultra in #4981
• add support for custom headers for LLM Generic OpenAI by @timothycarambat in #4999

New Contributors

• @j0rDy made their first contribution in #4894
• @vigorous-ride made their first contribution in #4964

Full Changelog: v1.10.0...v1.11.0

Highlighted Changes

AnythingLLM Desktop Assistant is live!

Now, AnythingLLM Desktop is a drop-in replacement for paid tools like Granola, Otter, Fireflies, and more.

• Runs entirely on your device, can record meetings without joining or summarize arbitrary files
• Powered by NVIDIA Parakeet + AnythingLLM's on-device orchastration
• Can call any agent tool, MCP, or anything else you already use with AnythingLLM!
• Custom summary templates, chat with the transcript, and even speaker identification.
• "Joined Meeting" Desktop notification to start a new recording with a click. For any meeting software (Zoom, Slack, Discord, Teams, etc)
• No rate limits, usage caps, or restrictions

AnythingLLM Mobile is live on Google Play

The Android AnythingLLM Mobile App is live on Google Play now. This syncs with both Cloud/Self-hosted and Desktop versions of AnythingLLM.

Notable other changes

• Removed onboarding "Create workspace" page -> goes straight to home now with new workspace in user native language
• Refactored Workspace file picker to be more performant
• Migrated Azure OpenAI to unified v1 api with full agent support
• Fixed Pagination bug in paperless-ngx
• Fixed issue where the undocumented YouTube API changed and broke the YT scraper
• Implemented Cohere as an agent provider
• A bump of dependency bumps
• Fixed bug where XSLX files dragged and dropped into chat weren't "visible" to the model
• MCP fixes for paths on non-Windows machines
• Docker image bumps and patches for a healthy Scout score (B)
• Added Error Boundary to UI to prevent white-page crashes

What's Changed

• bump body-parser patch version by @timothycarambat in #4746
• Remove bcrypt from collector - not used by @timothycarambat in #4747
• jws bump by @timothycarambat in #4748
• Refactor WorkspaceFileRow component by @shatfield4 in #4740
• Fix stale user permissions in UI by refreshing user data on app load by @angelplusultra in #4751
• Docker base image upgrade to Ubuntu 24 by @timothycarambat in #4754
• Bump expressJS from 4.18.2 -> 4.21.2 by @timothycarambat in #4760
• Migrate Azure OpenAI Integration To v1 API | Enable Streaming for Reasoning Models in Azure OpenAI Basic Inference Provider by @angelplusultra in #4744
• Bump form-data version by @timothycarambat in #4761
• Lazy load Lancedb by @timothycarambat in #4764
• Migrate to bcryptjs by @angelplusultra in #4767
• Refactor frontend legacy JSON.parse with safeJsonParse by @shatfield4 in #4759
• Fix pagination bug in paperless-ngx data connector by @shatfield4 in #4757
• Fix Stale User Session with Proper fetch Error Handling by @angelplusultra in #4770
• Add Auth Token to Ollama Embedding Client by @angelplusultra in #4766
• Upgrade to Multer 2.0.0 by @angelplusultra in #4768
• Implement Global Error Boundary by @angelplusultra in #4765
• Feat/cohere agent implementation by @17ColinMiPerry in #4703
• Upgrade MCP SDK to Latest (1.24.3) by @angelplusultra in #4773
• fix(devcontainer): forward ports 3000/3001 by @lohas1107 in #4779
• 4601 log model on response by @timothycarambat in #4781
• Update Google Search Option Description To Reference Documentation For Rate Limits by @angelplusultra in #4789
• Refactor LLMPerformanceMonitor.measureStream() to Use Options Object Pattern by @angelplusultra in #4786
• fix unnecessary scrollbar in workspace general appearance settings tab by @17ColinMiPerry in #4791
• Add Eslint Config in /frontend by @angelplusultra in #4785
• Refactor localWhisper to use custom FFMPEGWrapper class by @shatfield4 in #4775
• Refactor Managed Services in "Data Handling & Privacy" Onboarding Step to Use Their Privacy Policy URL by @angelplusultra in #4790
• patch ESM import issue by @timothycarambat in #4819
• Upgrade YT Scraper by @timothycarambat in #4820
• fix: validate chat message input by @Ocheretovich in #4811
• patch AWS credential issue in docker context by @timothycarambat in #4842
• support AWS bedrock agents with streaming by @timothycarambat in #4850
• VectorDB class migration by @shatfield4 in #4787
• Make XLSX spreadsheets visible in chat by combining sheets by @shatfield4 in #4847
• Remove Workspace Creation Onboarding Page by @shatfield4 in #4823
• Improved DMR support by @timothycarambat in #4863
• AnythingLLM Mobile live by @timothycarambat in #4864
• Adjust fix path to use ESM import by @timothycarambat in #4867
• Extract Model Table to component by @timothycarambat in #4871
• feat(i18n): add Czech (cs) language translation to AnythingLLM by @vladaman in #4874
• Docker model runner download from UI by @timothycarambat in #4884

New Contributors

• @lohas1107 made their first contribution in #4779
• @Ocheretovich made their first contribution in #4811
• @vladaman made their first contribution in #4874

Full Changelog: v1.9.1...v1.10.0

Notable Changes

• PWA Support: #4664
• Enable keyboard nav of slash commands with arrow keys on mount #4543
• Add ability to auto-handle YT video URLs in uploader & chat #4547
• Paperless ngx data connector by @shatfield4 in #4121
• Add SerpApi web search by @alexbarron in #4623
• Z.ai LLM & agent provider by @shatfield4 in #4573

What's Changed

• Fix KoboldCPP agent provider max tokens by @shatfield4 in #4519
• Enable Microsoft Foundry Local for workspace agent provider by @shatfield4 in #4539
• Enable keyboard nav of slash commands with arrow keys on mount by @timothycarambat in #4543
• Add ability to auto-handle YT video URLs in uploader & chat by @timothycarambat in #4547
• fix(server): correct Qdrant batching logic for large uploads by @aeromech-1 in #4545
• Agent workspace system prompt with variable expansion by @shatfield4 in #4526
• Refactor DefaultChatContainer To Display A Simple Welcome Message by @angelplusultra in #4542
• Refactor loadYouTubeTranscript() to include YouTube Video Metadata in Content When parseOnly is true by @angelplusultra in #4552
• Patch agent thoughts UI bug by @shatfield4 in #4549
• feat: support northflank deployment by @AdeboyeDN in #4570
• Install patch-package in /collector and Apply Patch to Fix EPub Upload Bug by @angelplusultra in #4630
• Failover invalid vector db identifier to lanceDB by @timothycarambat in #4661
• fix: unhelpful error message for invite link user creation by @naaa760 in #4621
• feat: Allow email addresses as usernames by @naaa760 in #4619
• Fix loop logic for fetchNextPage use in GitLabLoader by @timothycarambat in #4662
• add svelte rendering to markdown renderer by @timothycarambat in #4663
• Disable Prisma CLI Telemetry in Entrypoint Shell Script by @angelplusultra in #4609
• Support PWA installation by @timothycarambat in #4664
• Disable swagger docs endpoint with environment variable by @shatfield4 in #4665
• Paperless ngx data connector by @shatfield4 in #4121
• Implement full chat and @agent chat user indentificiation for OpenRouter by @timothycarambat in #4668
• fix: Use eval_duration for output TPS calculations in Ollama LLM provider by @jonathanortega2023 in #4568
• Add SerpApi web search by @alexbarron in #4623
• feature: Support for AWS Bedrock API Keys by @chetan in #4651
• Update Ollama AI Provider to Support Parsing "Thinking" Content From New Message Schema by @angelplusultra in #4587
• Z.ai LLM & agent provider by @shatfield4 in #4573
• Simplify cache condition for LMStudio and Ollama to prevent race condition by @timothycarambat in #4669
• Prompt caching for Anthropic LLM and Agent providers by @shatfield4 in #4488
• Fix undefined result in llm-instruction agent flow executor by @shatfield4 in #4676
• Custom Default System Prompt by @angelplusultra in #4487
• feat: add base64 document attachment support to API chat endpoints by @naaa760 in #4672
• fixed typos in General Settings: settings/embedding-preferences by @17ColinMiPerry in #4683
• Feat/add openrouter embedding models by @17ColinMiPerry in #4682
• Add batch embedding support for Ollama embedding provider by @shatfield4 in #4553
• Support Gitee AI(LLM Provider) by @fchange in #3361
• feat: Add SSL certificate bypass support for self-hosted Confluence instances by @naaa760 in #4219
• dev build with new epub2 build target and remove patch work by @timothycarambat in #4694
• docs(zh-TW): correct traditional Chinese translations in home section by @ascodeasice in #4702
• docs(i18n): Complete French translations by @raphael-bynome in #4697
• Store Chat Prompt Input Value in Local Storage by @angelplusultra in #4680
• Migrate to react router createBrowserRouter by @shatfield4 in #4693
• fixed empty scroll track in citations dropdown by @17ColinMiPerry in #4685
• Modify ChromaCloud provider to check base limitations of chroma cloud before upsert to prevent 422 by @timothycarambat in #4705
• Improve MCP functionality by @timothycarambat in #4709
• Patch dev pupeeteer crash for MacOS 15 by @timothycarambat in #4713
• Fix broken citations in Zilliz Cloud by @shatfield4 in #4716
• Fix similarity score not showing in citation UI for Astra, Qdrant, Weaviate by @shatfield4 in #4718
• Dell model capability patch by @timothycarambat in #4719

New Contributors

• @aeromech-1 made their first contribution in #4545
• @AdeboyeDN made their first contribution in #4570
• @jonathanortega2023 made their first contribution in #4568
• @alexbarron made their first contribution in #4623
• @chetan made their first contribution in #4651
• @fchange made their first contribution in #3361
• @ascodeasice made their first contribution in #4702
• @raphael-bynome made their first contribution in #4697

Full Changelog: v1.9.0...v1.9.1

@agent Overhaul & streaming ⚡️️

When anythingllm first launched, the word "agent" was not in the vocabulary of the LLM world. Agents are quickly becoming the standard for building AI applications and also the core experience for interacting with LLMs.

For too long, due to the complexity of building agents, spotty tool call support, models that can't even use tools and more nerd stuff, we often had to settle an experience that was not really fun to use since 99% of the time you were just looking at loading spinners waiting for the response.

The new agent experience is now here

Streams tool calls and responses in real time (all providers, all models)
Agents can now real-time download and ingest files from the web (eg: link to PDF, excel, csv). Anything you would use a document can be read in real time by the agent from the web.

Upcoming:

• Agent real-time API calling without agent flows
• Agent image understanding
• Agent system prompt passthrough + user context awareness
• Realtime file searching cross-platform default skill

Notable Improvements: 🚀

• All models and providers now support agentic streaming
• Microsoft Foundry Local integration
• Ephemerally scrape/download any web-resource via agent or uploader

What's Changed

• Allow default users to reorder workspaces by @shatfield4 in #4292
• Export image support for JSON and JSONL by @shatfield4 in #4359
• Fix: missing edit icon for prompts by @17ColinMiPerry in #4344
• feat(i18n): add missing Portuguese (Brazil) translations by @beckeryuri in #4328
• feat: Implement CometAPI integration for chat completions and model m… by @tensornull in #4379
• Resize chat textarea on paste by @shatfield4 in #4369
• update save file agent text by @timothycarambat in #4389
• Added metadata parameter to document/upload, document/upload/{folderName}, and document/upload-link by @jstawski in #4342
• Add support for SIMPLE_SSO_NO_LOGIN_REDIRECT config setting by @timothycarambat in #4394
• patch folder name GET request response by @timothycarambat in #4395
• Add User-Agent header on the requests sent by Generic OpenAI providers. by @angelplusultra in #4393
• Report sources in API responses on finalized chunk by @timothycarambat in #4396
• Allow user to specify args for chromium process so they dont need SYS_ADMIN on container. by @timothycarambat in #4397
• API request delay for Generic OpenAI embedding engine by @chaserhkj in #4317
• Enhanced Chat Embed History View by @MateKristof in #4281
• Ignore hasOwnProperty linting errors by @shatfield4 in #4406
• Migrate OpenAI LLM provider to use Responses API by @shatfield4 in #4404
• Update the timeout value on all stream-timeout providers: by @timothycarambat in #4412
• [BUGFIX] Update Dell Pro AI Studio Default URL by @spencerbull in #4433
• Add PostgreSQL vector extension in createTableIfNotExists function by @angelplusultra in #4430
• fix: resolve Firefox search icon overlapping placeholder text by @naaa760 in #4390
• Refactor Class Name Logging by @angelplusultra in #4426
• Change incorrect notation of Weaviate to PG Vector in env.example by @angelplusultra in #4439
• Enable custom HTTP response timeout for ollama by @timothycarambat in #4448
• fix: youtube transcript collector not work well with non en or non asr caption by @AoiYamada in #4442
• Add HTTP request/response logging middleware for development mode by @angelplusultra in #4425
• Sanitize Metadata Before PG Vector Database Insertion by @angelplusultra in #4434
• New Default System Prompt Variables (User ID, Workspace ID, & Workspace Name) by @angelplusultra in #4414
• Apply renderer from chat widget history to workspace chats by @timothycarambat in #4456
• Patch OpenAI metrics by @timothycarambat in #4458
• fix(uiux): correct typo in System Prompt description text by @vansh2408 in #4461
• Enable real-time agent tool call streaming for all providers by @timothycarambat in #4279
• Add stream options to Gemini LLM for usage tracking by @angelplusultra in #4466
• Fetch, Parse, and Create Documents for Statically Hosted Files by @angelplusultra in #4398
• Migrate OpenAI Agent to use ResponsesAPI by @timothycarambat in #4467
• Microsoft Foundry Local LLM provider & agent provider by @shatfield4 in #4435
• Model context limit auto-detection for LM Studio and Ollama LLM Providers by @shatfield4 in #4468
• Sync models from remote for FireworksAI by @timothycarambat in #4475
• Render html optional by @timothycarambat in #4478
• Adding AnythingLLM Helm Chart by @sculley in #4484
• Reimplement Cohere models for basic chat by @timothycarambat in #4489
• Tooltips for workspace and threads by @timothycarambat in #4500
• Improve URL handler for collector processes by @timothycarambat in #4504
• Migrate gemini agents away from Untooled by @timothycarambat in #4505
• Update .gitignore by @jaynedoezy-web in #4507
• refactor: change naming - contextwarpper to authprovider #4510 by @Guru6163 in #4511
• fix label for chunk length setting by @timothycarambat in #4515
• Fix: File pulling fails with uppercase URL characters by @angelplusultra in #4516

New Contributors

• @beckeryuri made their first contribution in #4328
• @tensornull made their first contribution in #4379
• @jstawski made their first contribution in #4342
• @angelplusultra made their first contribution in #4393
• @chaserhkj made their first contribution in #4317
• @MateKristof made their first contribution in #4281
• @spencerbull made their first contribution in #4433
• @AoiYamada made their first contribution in #4442
• @vansh2408 made their first contribution in #4461
• @sculley made their first contribution in #4484
• @jaynedoezy-web made their first contribution in #4507
• @Guru6163 made their first contribution in #4511

Full Changelog: v1.8.5...v1.9.0

AnythingLLM v1.8.5 is live

Notable Changes

Mobile support

Now, currently under Experimental features, you can connect the AnythingLLM Mobile App - Android Beta to your instance to seamlessly blend an on-device and off-device experience. Leverage your instance Agent Skills and flows all within a single unified interface!

Chat with documents has been overhauled

When we first built AnythingLLM, the average context window was 4K - hardly anything to fit a full document. So we decided to always be RAG first. This has its drawbacks since RAG is semantically dependent on asking questions about content in the document. This leads to poor results for "Summarize this document," only to be told by the document, "what are you talking about".

Well, now we have the best of both worlds. Documents are scoped to a workspace thread & user and we will attempt to use the full document text when possible and your model can support it. If you overflow this amount, we can then ask you to embed the document so you can unlock that long-term memory.

You can also easily manage and see your context window to remove files that are no longer relevant, but retain the conversation history.

You can also still embed files directly in the workspace file manager too :)

What's Changed

• Add model field parameter for generic OpenAI TTS by @timothycarambat in #4171
• Clear cached docs on text splitter changes by @shatfield4 in #4163
• Update BARE_METAL.md by @buraktopalyn in #4175
• run jest by @timothycarambat in #4177
• Moonshot AI LLM & agent provider by @shatfield4 in #4178
• Fix multimodal chats via openai compat api by @shatfield4 in #4135
• Native Embedder model selection (incl: Multilingual support) by @timothycarambat in #3835
• fix: correct some typos and grammatical errors in translation strings and index.jsx by @antonharder in #4195
• Modify the PostgreSQL SQL connector to support querying tables from schemas other than the default public schema. by @romanwu10 in #4202
• Fix app version metrics endpoint by @shatfield4 in #4209
• Update writeToServerDocuments to take config object by @timothycarambat in #4213
• [Chore]: sendCommand non positional call signature by @timothycarambat in #4218
• Update check for validFuncCall to only check required args and for undefined options by @timothycarambat in #4214
• fix: API export-chats endpoint function import error by @naaa760 in #4220
• STT append spoken text by @timothycarambat in #4216
• Mobile sync support by @timothycarambat in #4173
• Add custom JWT TTL by @timothycarambat in #4234
• Fix password hint text by @timothycarambat in #4235
• Update common.js Hebrew language update by @start-life in #4241
• docs(i18n): Complete missing Spanish translations by @timothycarambat in #4263
• feat: add Romanian translation by @dpvue in #4247
• Workspace Chat with documents overhaul by @timothycarambat in #4261
• feat: Add Exa as a Search Provider by @naaa760 in #4258
• Update common.js by @dpvue in #4278
• Chroma Cloud vector db provider by @shatfield4 in #4273
• Update Security UI to match all other Settings by @AshikaAnand12 in #4274

New Contributors

• @buraktopalyn made their first contribution in #4175
• @antonharder made their first contribution in #4195
• @romanwu10 made their first contribution in #4202
• @naaa760 made their first contribution in #4220
• @start-life made their first contribution in #4241
• @dpvue made their first contribution in #4247
• @AshikaAnand12 made their first contribution in #4274

Full Changelog: v1.8.4...v1.8.5

this is a minor patch update

Notable Changes

• Workspace & Thread searching now on the sidebar
• SQL Preflight connection validation (finally)
• Sticky codeblock headers while scrolling
• Codeblock max width to prevent long string overflows in the UI

What's Changed

• Handle BigInt in message response by @shatfield4 in #4110
• Configurable message limit for embed chat widget by @shatfield4 in #4114
• Add ability to search workspace and threads by @timothycarambat in #4120
• docs(i18n): Complete missing Chinese translations by @zz22558822 in #4123
• Fix chunking/snippet logs for clarity by @shatfield4 in #4129
• Sticky code block header by @shatfield4 in #4130
• docs(zh-cn): Update and improve Simplified Chinese translations by @ChenJiahao1 in #4137
• Sanitize youtube transcription file paths by @shatfield4 in #4148
• Add option to enable encryption on MSSQL Server db by @shatfield4 in #4134
• Migrate text size menu to embedded tooltip by @shatfield4 in #4147
• set max codeblock width to prevent overflows by @timothycarambat in #4157
• SQL preflight connection validation by @shatfield4 in #4150

New Contributors

• @zz22558822 made their first contribution in #4123
• @ChenJiahao1 made their first contribution in #4137

Full Changelog: v1.8.3...v1.8.4

AnythingLLM 1.8.3 is live!

News

Dont forget to sign up for the AnythingLLM mobile beta!

• Run small LLMs on device, full on-device RAG, agent tooling, and more. Fully private and on-device
• device-to-device sync between mobile and desktop clients over local networks

Notable Desktop Changes

• Built in Browser Tool You can now scrape private, authenticated, or otherwise gated context for collecting or use during agent web-scraping or agent flows!
• See the full changelog here

Notable updates and fixes

• Easily publish Agent Flows, Slash Commands, and System Prompts to the AnythingLLM Community Hub
• Patched YouTube transcript collector since it was failing due to Google API changes

What's Changed

• Flag to disable login UI and endpoints for credentialed auth by @timothycarambat in #3984
• Fix Typos in Comments and Documentation by @kilavvy in #3987
• Fix Typos in MetaGenerator.js Documentation Comments by @vtjl10 in #3993
• Fix Typo in Milvus Provider and Update Comment in Prisma Schema by @zeevick10 in #4003
• Fix typos in blog post example and save-file-browser plugin by @maximevtush in #4009
• Publish system prompts to hub by @shatfield4 in #3976
• Fix Typos in Comments and Documentation by @leopardracer in #4015
• Show app version in system settings sidebar by @timothycarambat in #4044
• Added missing German translations by @AlefAlfa in #4029
• Fix prompt, model, and temperature overrides in embed chat widget by @shatfield4 in #4036
• Publish agent flows to hub by @shatfield4 in #4001
• Publish slash commands to hub by @shatfield4 in #4019
• Added Estonian translation by @Tarkva in #4067
• Move show scrollbar option to chat settings by @shatfield4 in #4078
• Update common.js spellcheck for zh_TW by @martin0258 in #4074
• Add option to disable streaming via env for generic openai provider by @shatfield4 in #4079
• LMStudio 0 value embeddings patch by @shatfield4 in #4084
• Allow static defined prompt variables to be accessed by any authorized user by @shatfield4 in #4086
• Add hint for chats in query mode by @shatfield4 in #4076
• fix to support: feat2864 - using local credentials file with Amazon Bedrock by @nonatofabio in #3986
• Fix broken YT scraping with YT API by @bobbercheng in #4005
• Patch tokenizer race condition bug by @shatfield4 in #4091
• Polish translations by @timothycarambat in #4100
• feat: Add Korean (ko) locale translations by @ux-builder in #4096
• update typo in language key files by @timothycarambat in #4101
• Fix/drupal wiki (improve table & url handling) by @rexjohannes in #4097
• Fix score reporting for Milvus, Zilliz, and Pinecone vector databases by @shatfield4 in #4106

New Contributors

• @kilavvy made their first contribution in #3987
• @vtjl10 made their first contribution in #3993
• @zeevick10 made their first contribution in #4003
• @maximevtush made their first contribution in #4009
• @AlefAlfa made their first contribution in #4029
• @Tarkva made their first contribution in #4067
• @martin0258 made their first contribution in #4074
• @nonatofabio made their first contribution in #3986
• @bobbercheng made their first contribution in #4005

Full Changelog: v1.8.2...v1.8.3

AnythingLLM v1.8.2 is LIVE!

Other news

We were featured on stage at Microsoft Build (how cool???)

Notable Changes

Model swap in chat

You can now easily swap model or provider in the middle of chatting via cmd/Ctrl + Shift + L to show a tooltip menu to easily swap models without changing screens

System Prompt version tracking

Now, when you edit the system prompt of any workspace we locally store this information so you can easily refer to or restore to previous prompts that might work better for a given model or workspace.

What's Changed

• Fix/dom nesting warnings by @17ColinMiPerry in #3773
• Refactor AWS Bedrock Provider for Multi-modal Support & Correct Token Limits by @tristan-stahnke in #3714
• feat: support for iam roles for bedrock client by @dannysteenman in #2632
• feat: support fetch wikis for gitlab data connectors by @lcgash in #3271
• Updated README.md and README.zh_CN.md by @KPCOFGS in #3775
• Updated README.md and README.zh-CN.md by @KPCOFGS in #3776
• Enable workflow rule for package verification by @timothycarambat in #3778
• Update README.zh-CN.md by @KPCOFGS in #3780
• Chat prompt history menu by @shatfield4 in #3770
• minor UI patches by @timothycarambat in #3784
• Updated zh locale by @KPCOFGS in #3786
• locale change by @KPCOFGS in #3787
• Add Disable Spellcheck Option by @17ColinMiPerry in #3795
• Update:WorkspaceChat-ChatContainer-PromptInput by @maoguy in #3792
• updated locale by @KPCOFGS in #3796
• PGvector vector database support by @timothycarambat in #3788
• Fix MCP Start/Stop Toast Text by @17ColinMiPerry in #3803
• Fixed grammar and typos by @KPCOFGS in #3802
• Obsidian data connector by @shatfield4 in #3798
• Sponsors by @timothycarambat in #3816
• Fix typos by @omahs in #3826
• Update light mode designs for system prompt history by @shatfield4 in #3818
• Warn bad docker command by @timothycarambat in #3827
• Automatic Context window detection by @timothycarambat in #3817
• Improve UX for API keys and invitations by @shatfield4 in #3828
• Support Dell Pro AI Studio provider by @timothycarambat in #3829
• Optional summarization for web-scraping flow block by @shatfield4 in #3496
• 3416 file picker load times by @timothycarambat in #3833
• Respect Max Chunk Len set - Generic OpenAI Embedder connection by @timothycarambat in #3838
• Disable message send while content is embedding by @timothycarambat in #3832
• Updated Chinese locale by @KPCOFGS in #3852
• Filter malformed post-processed docs from file picker by @timothycarambat in #3853
• Update Brazilian Portuguese locale (pt_BR) by @JairoMarques in #3879
• Removed redundant scrollbar in search containers by @17ColinMiPerry in #3865
• Fix File attachments are not cleared on second send click when use @A… by @cnjsyyb in #3843
• Update citations UI by @shatfield4 in #3837
• Drupal invisible output with connector by @timothycarambat in #3878
• fixed drupal connector by @rexjohannes in #3893
• Feat/missing localizations by @JairoMarques in #3881
• Keyboard shortcut support by @AshAnand34 in #3890
• Handle invalid response bodies for ContextWindowFinder by @timothycarambat in #3896
• Add Latvian language by @timothycarambat in #3897
• Update styles of attachments uploaded in PromptInput by @shatfield4 in #3892
• Remove border around image upload component by @shatfield4 in #3900
• fix issue with files loading with watching enabled by @timothycarambat in #3930
• Update more Brazilian Portuguese locale (pt_BR) by @JairoMarques in #3923
• Fix/mcp environment inheritance by @PaoloC68 in #3909
• Add support for PUT and PATCH in agent flow builder API call block by @shatfield4 in #3937
• Add missing localizations by @JairoMarques in #3925
• Patch apiChatHandler to include agent chats by @shatfield4 in #3905
• New embedded chat widget UI/UX by @shatfield4 in #3899
• JSON Parser from agent flow blocks by @shatfield4 in #3863
• remove LLM Instruction inputVariable requirement by @timothycarambat in #3954
• ManageWorkspace styles update by @shatfield4 in #3948
• Direct output for agent flows by @shatfield4 in #3873
• Implement importing of agent flows from community hub by @shatfield4 in #3867
• Fix some french translations by @Stoufiler in #3959
• Fix text wrapping in directory tooltip by @shatfield4 in #3967
• Fix model map staleness behavior or fallback by @timothycarambat in #3971
• Enable UI/UX for model swapping in chat window by @timothycarambat in #3969
• Fix Typo in Documentation and Metadata Field Name by @leopardracer in #3979

New Contributors

• @17ColinMiPerry made their first contribution in #3773
• @tristan-stahnke made their first contribution in #3714
• @dannysteenman made their first contribution in #2632
• @maoguy made their first contribution in #3792
• @omahs made their first contribution in #3826
• @JairoMarques made their first contribution in #3879
• @cnjsyyb made their first contribution in #3843
• @rexjohannes made their first contribution in #3893
• @AshAnand34 made their first contribution in #3890
• @PaoloC68 made their first contribution in #3909
• @Stoufiler made their first contribution in #3959
• @leopardracer made their first contribution in #3979

Full Changelog: v1.8.1...v.1.8.2

What's Changed

• enable MCP tool use via @agent in API calls by @timothycarambat in #3683
• Add support for /v1/document/upload* endpoints to support auto-add to workspace by @timothycarambat in #3692
• Feature/drupalwiki collector by @timothycarambat in #3693
• Enable bypass of ip limitations via ENV in collector processing by @timothycarambat in #3652
• Translate main page landing by @timothycarambat in #3699
• Allow custom headers in upload-link endpoint by @shatfield4 in #3695
• Improve Japanese translation to maintain @agent by @suzuki-shm in #3704
• Update embed by @timothycarambat in #3728
• TTS auto-speak and STT auto-submit by @timothycarambat in #3696
• Update Bing Search API Azure Service copy - Unavailable by @thejjw in #3729
• Update light mode input color by @shatfield4 in #3734
• Replace @azure/openai with openai, and update openai to version 4.95.1 by @suzuki-shm in #3691
• Remove @google/generative-ai from AnythingLLM by @timothycarambat in #3740
• Update README.zh-CN.md by @KPCOFGS in #3748
• pt-BR translation by @thisisleobro in #3755
• #3707 added sse and streamable stransport support for mcp servers by @AlekseyKapustyanenko in #3747
• App wide antialiasing by @shatfield4 in #3752
• Add option to control KoboldCPP max response tokens by @shatfield4 in #3746
• handle no thoughts in container by @timothycarambat in #3758
• Remove drop shadow from CTAButton component by @shatfield4 in #3756
• Fix chat alignment borders by @shatfield4 in #3760
• Data connector CTA button styles by @shatfield4 in #3759
• Add gemini-embedding-exp-03-07 model support by @sdarioz in #3767
• Filter image models from OpenAI by @shatfield4 in #3768
• more zh translations by @timothycarambat in #3769

New Contributors

• @thejjw made their first contribution in #3729
• @thisisleobro made their first contribution in #3755
• @AlekseyKapustyanenko made their first contribution in #3747
• @sdarioz made their first contribution in #3767

Desktop App Changelog

https://docs.anythingllm.com/changelog/v1.8.1

Full Changelog: v1.8.0...v1.8.1