Cross-platform tool and drivers for liquid coolers and other devices
Notice: please check out our open invitation for new team members.
$ liquidctl list
Device #0: Corsair Vengeance RGB DIMM2
Device #1: Corsair Vengeance RGB DIMM4
Device #2: NZXT Smart Device (V1)
Device #3: NZXT Kraken X (X42, X52, X62 or X72)
# liquidctl initialize all
NZXT Smart Device (V1)
├── Firmware version 1.7
├── LED accessories 2
├── LED accessory type HUE+ Strip
└── LED count (total) 20
NZXT Kraken X (X42, X52, X62 or X72)
└── Firmware version 6.2
# liquidctl status
NZXT Smart Device (V1)
├── Fan 1 speed 1499 rpm
├── Fan 1 voltage 11.91 V
├── Fan 1 current 0.05 A
├── Fan 1 control mode PWM
├── Fan 2 [...]
├── Fan 3 [...]
└── Noise level 61 dB
NZXT Kraken X (X42, X52, X62 or X72)
├── Liquid temperature 34.7 °C
├── Fan speed 798 rpm
└── Pump speed 2268 rpm
# liquidctl status --match vengeance --unsafe=smbus,vengeance_rgb
Corsair Vengeance RGB DIMM2
└── Temperature 37.5 °C
Corsair Vengeance RGB DIMM4
└── Temperature 37.8 °C
# liquidctl --match kraken set fan speed 20 30 30 50 34 80 40 90 50 100
# liquidctl --match kraken set pump speed 70
# liquidctl --match kraken set sync color fixed 0080ff
# liquidctl --match "smart device" set led color moving-alternating "hsv(30,98,100)" "hsv(30,98,10)" --speed slower
- Supported devices
- Installation
- The command-line interface
- Using liquidctl in other programs and scripts
- Automation and running at boot
- Troubleshooting
- Additional documentation
- License
- Related projects
The following devices are supported by liquidctl. In the table, MRLV stands for the minimum recommended liquidctl version. The linked documents contain specific usage instructions and other useful information.
B Known to be currently broken in at least one major way.
L Requires the --legacy-690lc flag.
U Requires --unsafe features.
Z Requires replacing the device driver on Windows.
a Architecture-specific limitations.
h Can leverage hwmon driver.
p Only partially supported.
x Only supported on Linux.
The following sections cover the various methods to set up liquidctl.
A considerable number of Linux distributions already package liquidctl, generally at fairly recent versions.
# Alpine
sudo apk add liquidctl
# Arch/Artix/[Manjaro]/Parabola
sudo pacman -S liquidctl
# Fedora
sudo dnf install liquidctl
# Manjaro
sudo pamac install liquidctl
# Nix
nix-env -iA nixos.liquidctlliquidctl is also available in some non-official/community-based repositories, as well as, at older versions, for more distributions. Repology shows more information about the packaging status in various distributions.
For macOS, liquidctl is available on Homebrew, generally at the most recent version. It is also easy to install the latest development snapshot from the official source code repository.
# latest stable version
brew install liquidctl
# or latest development snapshot from the official source code repository
brew install liquidctl --HEADOn FreeBSD and DragonFly BSD, liquidctl is maintained in the Ports Collections, and is available as a pre-built binary package.
pkg install py39-liquidctl
Warning: on systems that still default to Python 2, replace python
with python3.
Changed in 1.9.0: liquidctl now uses a PEP 517 build system.
liquidctl can be manually installed from the Python Package Index (PyPI), or directly from the source code repository.
In order to manually install it, certain system-level dependencies must be satisfied first. In some cases it may also be preferable to use the Python libraries already provided by the operating system.
On Linux, the following dependencies are required at runtime (common package names are listed in parenthesis):
- Python 3.8 or later (python3, python)
- pkg_resources Python package (python3-setuptools, python3-pkg-resources, python-setuptools)
- PyUSB (python3-pyusb, python3-usb, python-pyusb)
- colorlog (python3-colorlog, python-colorlog)
- crcmod 1.7 (python3-crcmod, python-crcmod)
- cython-hidapi (python3-hidapi, python3-hid, python-hidapi)
- docopt (python3-docopt, python-docopt)
- pillow (python-pillow, python3-pil)
- smbus Python package (python3-i2c-tools, python3-smbus, i2c-tools)
- LibUSB 1.0 (libusb-1.0, libusb-1.0-0, libusbx)
Additionally, to build, install and test liquidctl, the following are also needed:
- setuptools_scm Python package (python3-setuptools-scm, python3-setuptools_scm, python-setuptools-scm)
- pip (optional) (python3-pip, python-pip)
- pytest (optional) (python3-pytest, pytest, python-pytest)
On macOS, Python (3.8 or later) and LibUSB 1.0 must be installed beforehand.
brew install python libusb
On Windows, Python (3.8 or later) must be installed beforehand, which can be
done from the official website. It is recommended to select the
option to add python and other tools to the PATH.
A LibUSB 1.0 DLL is also necessary, but it will generally be provided
automatically by liquidctl. In case that's not possible, and a USB "No backend
available" error is shown, the suitable DLL from an official LibUSB release
should be copied into C:\Windows\System32\. The DLL must match your Python
installation: in most cases it will be latest VS build for x64 in the archive
from LibUSB (e.g. VS2015-x64/dll/libusb-1.0.dll).
Additionally, products that are not Human Interface Devices (HIDs), or that do
not use the Microsoft HID Driver, require a libusb-compatible driver (these are
listed in Supported devices with a Z note). In most cases of these cases
the Microsoft WinUSB driver is recommended, and it can easily be set up for a
device using Zadig: open Zadig, select your device from the dropdown list
and, finally, click "Replace Driver".
Warning: replacing the driver for a device where that is not necessary will
likely cause it to become inaccessible from liquidctl.
Changed in 1.9.0: a LibUSB 1.0 DLL is now provided by libusb-package, provided
there are suitable wheels available at the time of installation.
Setting up a virtual environment is an optional step. Even so, installing Python packages directly in the global environment is not generally advised.
Instead, it is usual to first set up a virtual environment:
# create virtual enviroment at <path>
python -m venv <path>Once set up, the virtual environment can be activated on the current shell
(more information in the official documentation).
Alternatively, the virtual environment can also be used directly, without
activation, by prefixing all python invocations with the environment's bin
directory.
# Linux/macOS/BSDs (POSIX)
<path>/bin/python [arguments]
# Windows
<path>\Scripts\python [arguments]pip can be used to install liquidctl from the Python Package Index (PyPI). This will also install the necessary Python libraries.
# the latest stable version
python -m pip install liquidctl
# a specific version (e.g. 1.13.0)
python -m pip install liquidctl==1.13.0If git is installed, pip can also install the latest snapshot of the official liquidctl source code repository on GitHub.
# the latest snapshot of the official source code repository (requires git)
python -m pip install git+https://github.com/liquidctl/liquidctl#egg=liquidctlTo set up rootless access to devices on Linux and BSDs, install documentation and completions, or to learn more about auxiliary scripts, continue reading for the next few sections.
Access permissions are not a concern on platforms like macOS or Windows, where unprivileged access is already allowed by default. However, devices are not generally accessible by unprivileged users on Linux, FreeBSD or DragonFly BSD.
For Linux, we provide a set of udev rules in 71-liquidctl.rules that can be
used to allow unprivileged read and write access to all devices supported by
liquidctl. These rules are generally already included in downstream Linux
packages of liquidctl.
Alternatively, sudo, doas and similar mechanisms can be used to invoke
liquidctl as the super user, on both Linux and BSDs.
Other files and tools are included in the source tree, which may be of use in certain scenarios:
- liquidctl(8) man page;
- completions for the liquidctl CLI in Bash;
- host-based automatic fan/pump speed control;
- send liquidctl data to HWiNFO;
- and more....
Changed in 1.9.0: liquidctl now uses a PEP 517 build system.
When working on the project itself, it is sometimes useful to set up a local development environment, making it possible to directly run the CLI and the test suite, without first building and installing a local package.
For this, start by installing git and any system-level dependencies mentioned in Manual installation. Then, clone the repository and change into the created directory:
git clone https://github.com/liquidctl/liquidctl
cd liquidctl
Optionally, set up a virtual environment.
Finally, if the necessary Python build, test and runtime libraries are not already installed on the environment (virtual or global), manually install them:
python -m pip install --upgrade pip setuptools setuptools_scm wheel
python -m pip install --upgrade colorlog crcmod==1.7 docopt hidapi pillow pytest pyusb
python -m pip install --upgrade "libusb-package; sys_platform == 'win32' or sys_platform == 'cygwin'"
python -m pip install --upgrade "smbus; sys_platform == 'linux'"
python -m pip install --upgrade "winusbcdc>=1.5; sys_platform == 'win32'"
At this point, the environment is set up. To run the test suite, execute:
python -m pytest
To run the CLI directly, without building and installing a local package, execute:
python -m liquidctl [arguments]
And to install liquidctl into the environment:
python -m pip install .
The complete list of commands and options can be found in liquidctl --help and in the man page, but the following topics cover the most common operations.
Brackets [ ], parenthesis ( ), less than/greater than < > and ellipsis ... are used to describe, respectively, optional, required, positional and repeating elements. Example commands are prefixed with a number sign #, which also serves to indicate that on Linux root permissions (or suitable udev rules) may be required.
The --verbose option will print some extra information, like automatically made adjustments to user-provided settings. And if there is a problem, the --debug flag will make liquidctl output more information to help identify its cause; be sure to include this when opening a new issue.
Note: in addition to --debug, setting the PYUSB_DEBUG=debug and LIBUSB_DEBUG=4 environment variables can be helpful with problems suspected to relate to PyUSB or LibUSB.
A good place to start is to ask liquidctl to list all recognized devices.
$ liquidctl list
Device #0: NZXT Smart Device (V1)
Device #1: NZXT Kraken X (X42, X52, X62 or X72)
In case more than one supported device is found, one them can be selected with --match <substring>, where <substring> matches part of the desired device's description using a case insensitive comparison.
$ liquidctl --match kraken list
Result #0: NZXT Kraken X (X42, X52, X62 or X72)
More device properties can be show by passing --verbose to liquidctl list. Any of those can also be used to select a particular product.
$ liquidctl --bus hid --address /dev/hidraw4 list
Result #0: NZXT Smart Device (V1)
$ liquidctl --serial 1234567890 list
Result #0: NZXT Kraken X (X42, X52, X62 or X72)
Ambiguities for any given filter can be solved with --pick <number>.
Devices will usually need to be initialized before they can be used, though each device has its own requirements and limitations. This and other information specific to a particular device will appear on the documentation linked from the Supported devices section.
Devices can be initialized individually or all at once.
# liquidctl [options] initialize [all]
Most devices provide some status information, like fan speeds and liquid temperatures. This can be queried for all devices or using the filtering methods mentioned before.
# liquidctl [options] status
Fan and pump speeds can be set to fixed values or, if the device supports them, custom profiles. The specific documentation for each device will list the available modes, as well as which sensor is used for custom profiles. In general, liquid coolers only support custom profiles that are based on the internal liquid temperature probe.
# liquidctl [options] set <channel> speed (<temperature> <percentage>) ...
# liquidctl [options] set <channel> speed <percentage>
Lighting is controlled in a similar fashion. The specific documentation for each device will list the available channels, modes and additional options.
# liquidctl [options] set <channel> color <mode> [<color>] ...
When configuring lighting effects, colors can be specified in different representations and formats:
- as an implicit hexadecimal RGB triple, either with or without the
0xprefix: e.g.ff7f3f - as an explicit RGB triple: e.g.
rgb(255, 127, 63) - as a HSV (hue‑saturation‑value) triple: e.g.
hsv(20, 75, 100)- hue ∊ [0, 360] (degrees); saturation, value ∊ [0, 100] (percent)
- note: this is sometimes called HSB (hue‑saturation‑brightness)
- as a HSL (hue‑saturation‑lightness) triple: e.g.
hsl(20, 100, 62)- hue ∊ [0, 360] (degrees); saturation, lightness ∊ [0, 100] (percent)
Color arguments containing spaces, parenthesis or commas need to be quoted, as these characters can have special meaning on the command-line; the easiest way to do this on all supported platforms is with double quotes.
# liquidctl --match kraken set ring color fading "hsv(0,80,100)" "hsv(180,80,100)"
On Linux it is also possible to use single-quotes and \(, \), \ escape sequences.
The liquidctl driver APIs can be used to build Python programs that monitor or control the devices, and offer features beyond the ones provided by the CLI.
The APIs are documented, and this documentation can be accessed through
pydoc, or directly read from the source files.
from liquidctl import find_liquidctl_devices
# Find all connected and supported devices.
devices = find_liquidctl_devices()
for dev in devices:
# Connect to the device. In this example we use a context manager, but
# the connection can also be manually managed. The context manager
# automatically calls `disconnect`; when managing the connection
# manually, `disconnect` must eventually be called, even if an
# exception is raised.
with dev.connect():
print(f'{dev.description} at {dev.bus}:{dev.address}:')
# Devices should be initialized after every boot. In this example
# we assume that this has not been done before.
print('- initialize')
init_status = dev.initialize()
# Print all data returned by `initialize`.
if init_status:
for key, value, unit in init_status:
print(f'- {key}: {value} {unit}')
# Get regular status information from the device.
status = dev.get_status()
# Print all data returned by `get_status`.
print('- get status')
for key, value, unit in status:
print(f'- {key}: {value} {unit}')
# For a particular device, set the pump LEDs to red.
if 'Kraken' in dev.description:
print('- set pump to radical red')
radical_red = [0xff, 0x35, 0x5e]
dev.set_color(channel='pump', mode='fixed', colors=[radical_red])More examples can be found in the scripts in extra/.
In addition to the APIs, the liquidctl CLI is friendly to scripting: errors
cause it to exit with non-zero codes and only functional output goes to
stdout, everything else (error messages, warnings and other auxiliary
information) going to stderr.
The list, initialize and status commands also support a --json flag to
switch the output to JSON, a more convenient format for machines and scripts.
In --json mode, setting LANG=C on the environment causes non-ASCII
characters to be escaped.
# liquidctl --match kraken list --json | jq
[
{
"description": "NZXT Kraken X (X42, X52, X62 or X72)",
"vendor_id": 7793,
"product_id": 5902,
"release_number": 512,
"serial_number": "49874481333",
"bus": "hid",
"address": "/dev/hidraw3",
"port": null,
"driver": "Kraken2",
},
...
]
# liquidctl --match kraken status --json | jq
[
{
"bus": "hid",
"address": "/dev/hidraw3",
"description": "NZXT Kraken X (X42, X52, X62 or X72)",
"status": [
{
"key": "Liquid temperature",
"value": 30.1,
"unit": "°C"
},
{
"key": "Fan speed",
"value": 1014,
"unit": "rpm"
},
...
]
},
...
]
Note that the examples above pipe the output to jq, as the original output
has no line breaks or indentation. An alternative to jq is to use python -m json.tool, which is already included in standard Python
distributions.
Finally, the stability of both the APIs and the CLI commands is documented in our stability guarantee. In particular, the specific keys, values and units returned by the commands above, as well as their API equivalents, are subject to changes. Consumers should verify that the returned data matches their expectations, and react accordingly.
In most cases you will want to automatically apply your settings when the system boots. Generally a simple script or a basic service is enough, and some specifics about this are given in the following sections.
For even more flexibility, you can also write a Python program that calls the driver APIs directly.
On systems running Linux and systemd a service unit can be used to configure liquidctl devices. A simple example is provided bellow, which you can edit to match your preferences. Save it to /etc/systemd/system/liquidcfg.service.
[Unit]
Description=AIO startup service
[Service]
Type=oneshot
ExecStart=liquidctl initialize all
ExecStart=liquidctl --match kraken set pump speed 90
ExecStart=liquidctl --match kraken set fan speed 20 30 30 50 34 80 40 90 50 100
ExecStart=liquidctl --match "smart device" set sync speed 55
ExecStart=liquidctl --match kraken set sync color fading 350017 ff2608
[Install]
WantedBy=default.target
After reloading the configuration, the new unit can be started manually or set to automatically run during boot using standard systemd tools.
# systemctl daemon-reload
# systemctl start liquidcfg
# systemctl enable liquidcfg
A slightly more complex example can be seen at jonasmalacofilho/dotfiles, which includes dynamic adjustments of the lighting depending on the time of day.
If necessary, it is also possible to have the service unit explicitly wait for the device to be available: see making systemd units wait for devices.
The configuration of devices can be automated by writing a batch file and setting up a new task for (every) login using Windows Task Scheduler. The batch file can be really simple and only needs to contain the invocations of liquidctl that would otherwise be done manually.
liquidctl initialize all
liquidctl --match kraken set pump speed 90
liquidctl --match kraken set fan speed 20 30 30 50 34 80 40 90 50 100
liquidctl --match "smart device" set sync speed 55
liquidctl --match kraken set sync color fading 350017 ff2608Make sure that liquidctl is available in the context where the batch file will run: in short, liquidctl --version should work within a normal Command Prompt window.
You may need to install Python with the option to set the PATH variable enabled, or manually add the necessary folders to the PATH.
A slightly more complex example can be seen in issue #14 ("Can I autostart liquidctl on Windows?"), that uses the LEDs to convey progress or eventual errors. Chris' guide on Replacing NZXT’s CAM software on Windows for Kraken is also a good read.
As an alternative to using Task Scheduler, the batch file can simply be placed in the startup folder; you can run shell:startup to find out where that is.
You can follow either or both of the guides below to automatically configure your devices during login or after waking from sleep. The guides are hosted on tonymacx86:
- This guide is for controllers that lose their state during sleep (e.g. Gigabyte RGB Fusion 2.0) and need to be reinitialized after wake-from-sleep. This guide uses Automator to initialize supported devices at login, and sleepwatcher to initialize supported devices after wake-from-sleep.
- This guide is for controllers that do not lose their state during sleep (e.g. ASUS Aura LED). This driver uses the launchctl method to initialize supported devices at login.
This is likely caused by having replaced the standard driver of a USB HID. If the device in question is not marked in Supported devices as requiring a special driver, try uninstalling the custom driver.
First, check that the liquidctl version (liquidctl --version) you're running supports the device in question. An overview of supported devices and minimum recommended versions for them are available in Supported devices. Support for new devices can happen to be available only in Git for some time.
If the device is supported, this is usually caused by having an unexpected kernel driver bound to a USB HID. In most cases this is the result of having used a program that accessed the device (directly or indirectly) via libusb-1.0, but failed to reattach the original driver before terminating.
This can be temporarily solved by manually rebinding the device to the kernel usbhid driver. Replace <bus> and <port> with the correct values from lsusb -vt (also assumes there is only HID interface, adjust if necessary):
echo '<bus>-<port>:1.0' | sudo tee /sys/bus/usb/drivers/usbhid/bind
A more permanent solution is to politely ask the authors of the program that is responsible for leaving the kernel driver detached to use libusb_attach_kernel_driver or libusb_set_auto_detach_kernel_driver.
These errors are usually caused by a lack of permission to access the device. On Linux distros that normally requires root privileges.
Alternatively to running liquidctl as root (or with sudo), you can install the udev rules provided in extra/linux/71-liquidctl.rules to allow unprivileged access to the devices supported by liquidctl.
If your problem is not listed here, try searching the issues. If no issue matches your problem, you still need help, or you have found a bug, please open one.
When commenting on an issue, please describe the problem in as much detail as possible. List your operating system and the specific devices you own.
Also include the arguments and output of all relevant/failing liquidctl commands, using the --debug option to enable additional debug information.
Be sure to browse docs/ for additional documentation, and extra/ for some example scripts and other possibly useful things.
You are also encouraged to contribute to the documentation and to these examples, including adding new files that cover your specific use cases or solutions.
Copyright 2018–2023 Jonas Malaco, Marshall Asch, CaseySJ, Tom Frey, Andrew Robertson, ParkerMc, Aleksa Savic, Shady Nawara and contributors
Some modules also incorporate or use as reference work by leaty, Ksenija Stanojevic, Alexander Tong, Jens Neumaier, Kristóf Jakab, Sean Nelson, Chris Griffith, notaz, realies and Thomas Pircher. This is mentioned in the module docstring, along with appropriate additional copyright notices.
This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but without any warranty; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with this program. If not, see https://www.gnu.org/licenses/.
Sibling project of Linux kernel hwmon drivers for devices supported by liquidctl.
Graphical interface to monitor and control cooling devices supported by liquidctl.
Graphical interface to control many different types of RGB devices.