Getting Started with LM Studio for Local LLMs

LM Studio is one of the most accessible ways to run local language models on a personal machine. It bundles model discovery, download, inference, and an OpenAI-compatible API into a single desktop application, making it an attractive option for people who want local AI without the command-line overhead.

What you need before installing

LM Studio runs on macOS, Windows, and Linux. It works best on systems with a dedicated GPU and at least 16 GB of system memory, but smaller quantised models can run on machines with less hardware. Download the installer from the official website and run through the standard setup steps.

For a broader view of hardware requirements, read Best Hardware for Self-Hosted AI.

Browse and download models

The built-in model browser connects to Hugging Face and lets you search, filter, and download models directly. Start with a popular quantised model in the 7B to 13B parameter range — these offer a good balance of quality and performance on consumer hardware.

Understand GGUF format

LM Studio uses GGUF-format models, which are pre-quantised for efficient CPU and GPU inference. The format keeps model files smaller and loading times faster compared to raw model weights.

Run your first inference session

Once the model is downloaded, select it from the model list, adjust the context length if your GPU memory allows, and start chatting. The interface shows token generation speed, memory usage, and model metadata so you can track performance.

If you prefer a web-based interface, compare the experience in Open WebUI vs AnythingLLM.

Enable the local API server

LM Studio includes a local HTTP server that exposes an OpenAI-compatible endpoint. This is useful for connecting LM Studio to other tools, automation workflows, and chat interfaces. Enable it from the server tab, note the port, and point your tools to `http://localhost:1234/v1`.

For workflow automation ideas, see Build Your Own AI Assistant with n8n.

Switch models without restarting

One of LM Studio's strengths is the ability to load and unload models on the fly. You can switch between a fast coding model and a larger reasoning model within the same session, keeping the API server running throughout.

Conclusion

LM Studio lowers the barrier for local LLM experimentation. Its graphical interface, built-in model browser, and OpenAI-compatible API make it a strong choice for beginners and intermediate users who prefer a desktop experience over terminal-based workflows.

FAQ

Is LM Studio free?

Yes. The application is free to download and use with any compatible model from Hugging Face.

Can LM Studio use multiple GPUs?

It supports offloading layers across GPU and CPU, but full multi-GPU support depends on the backend and model format.

Does LM Studio support vision models?

Vision-language models are supported through the same interface when the model is compatible with the underlying inference engine.