lmstudio/llmster-preview repository overview

_    _    __  __ ___ _____ ___ ___
| |  | |  |  \/  / __|_   _| __| _ \
| |__| |__| |\/| \__ \ | | | _||   /
|____|____|_|  |_|___/ |_| |___|_|_\
headless lmstudio • docker image

⁠llmster

[Technical Preview] llmster is the headless version of LM Studio⁠. This image provides a way to run the LM Studio daemon (llmster) without requiring specialized GPU hardware. Suitable for CI jobs, automations, and testing on commodity CPUs.

Currently, this image is in preview and has CPU-only support on x86 systems

⁠Quick Start

Pull the image:

docker pull lmstudio/llmster-preview:cpu

Run the container:

docker run -d -p 1234:1234 lmstudio/llmster-preview:cpu

The daemon API server will be accessible at http://<container-host>:1234. Find out more details in the documentation⁠

<container-host> depends on your context:

• If accessing from the same machine: use localhost
• If accessing from another machine: use the IP address or hostname of the machine running Docker
• If accessing from another container in the same Docker network: use the container name or service name (in Docker Compose)

⁠Usage

⁠lms CLI

The recommended way to use this docker container is through the lms CLI tool which comes pre-installed. You can interact with the daemon in two ways:

1. From your host machine: Use lms <COMMAND> --host <container-host> --port 1234 to download, load and run various models.
2. From within the container: Use docker exec -it <container-id> bash and then use lms directly.

⁠CLI Examples

Here are a few examples of using the CLI. If using from the host machine, remember to include --host <container-host> --port 1234 with each command. Find out more in our CLI documentation⁠.

⁠Download a model

lms get google/gemma-3-1b

⁠Load a model

lms load google/gemma-3-1b

⁠Chat with a model

lms chat --prompt "Hello, how does Docker work? Explain in 3 sentences or less"

⁠RESTful API

This image has an internal REST API server which can be interacted through the exposed 1234 port. For more information, including API documentation and advanced usage, visit LM Studio API Documentation⁠.