Supercharge your AI assistant by connecting it to local Large Language Models (LLMs) running on your own computer via Ollama, removing the need for internet access or cloud APIs. This Python-based server exposes your self-hosted Ollama models as a suite of eight distinct tools, allowing AI assistants like Windsurf, VS Code, and Claude Desktop to leverage local LLMs for chat, text generation, embeddings, model management, and more. It provides a universal, lightweight, and type-safe solution for robust local AI integration.
Características Principales
018 MCP tools for full Ollama SDK access
02Hot-swap architecture for auto-discovery of new tools
03Type-safe with Pydantic models
04Lightweight with minimal dependencies and fast startup
05Universal compatibility with any MCP client
061 GitHub stars
Casos de Uso
01Enable VS Code or Claude Desktop to chat and generate text using self-hosted Ollama models
02Integrate local LLMs with AI assistants like Windsurf for offline AI capabilities
03Manage and interact with local Ollama models (pull, delete, list, inspect) directly from an AI assistant
