LMStudio-MCP
LMStudio-MCP is a Model Control Protocol server that integrates Claude with locally running LM models via LM Studio, allowing health checks, model listings, and text generation. It serves as a bridge to combine local model processing with Claude's interface.
What are the prerequisites for using LMStudio-MCP?
You need Python 3.7+, LM Studio installed and running locally with a model loaded, Claude with MCP access, and the required Python packages.
How do I start the LMStudio-MCP server?
Ensure LM Studio is running on port 1234, load a model, and run the server using python lmstudio_bridge.py if running locally.
What should I do if I encounter API connection issues?
Ensure LM Studio is running with a loaded model, check the server is on port 1234, verify firewall settings, and try using '127.0.0.1' instead of 'localhost'.
Are there any known limitations with LMStudio-MCP?
Some models may have compatibility issues, and the bridge uses only OpenAI-compatible API endpoints. Model responses depend on the capabilities of the loaded model.
How can I troubleshoot model compatibility issues?
Try different parameter values for problematic models or switch to a more compatible model if issues persist.