lmstudio-mcp

# LMStudio API MCP Server This is a FastMCP server that provides an interface to interact with LMStudio's local language models. It allows you to query and interact with models loaded in LMStudio through a simple API. ## Features - List all loaded models in LMStudio - Send prompts to specific models with custom system prompts - Simple API interface for model interactions ## Prerequisites - Python 3.x - LMStudio installed and running - FastMCP package installed - LMStudio Python package installed ## Usage The server provides two main endpoints: ### 1. Get Available Models ```python get_lmstudio_models() -> list[str] ``` Returns a list of all models currently loaded in LMStudio. Each model entry contains information about the model including its display name, which is used to identify the model in subsequent API calls. ### 2. Query a Model ```python ask_lmstudio_model(model: str, system_prompt: str, prompt: str) -> str ``` Sends a prompt to a specific model with a custom system prompt. Parameters: - `model`: The display name of the model to use - `system_prompt`: The system prompt to set the context - `prompt`: The actual prompt to send to the model ## Example ```python # Get list of available models models = get_lmstudio_models() # Ask a question to a specific model response = ask_lmstudio_model( model="Your Model Name", system_prompt="You are a helpful assistant.", prompt="What is the capital of France?" ) ``` ## Running the Server To start the MCP server, simply run: ```bash python lmstudio_api.py ``` ## Error Handling The server includes basic error handling: - Returns "Model not found" if the specified model doesn't exist - Returns "No response from model" if the model fails to generate a response ## Note Make sure LMStudio is running and the desired models are loaded before using this API.