ollama-mcp

# Ollama MCP Client A Python-based client implementation for connecting Ollama large language models and MCP (Model Context Protocol) servers, enabling seamless interaction between models and external tools. ## Features - Integrated Ollama large language model (defaults to qwen3:14b) - Supports SSE (Server-Sent Events) asynchronous communication - Extensible tool calling system - Interactive command-line interface - Simple tool format conversion (Spring AI's tools_list format ⟷ Ollama format) ## Quick Start ### Environment Requirements - Python 3.12+ - Ollama CLI (refer to [Ollama official website](https://ollama.com/) for installation instructions) - java 21+ ### Installation ```bash git clone https://github.com/Shlysz/ollama-mcp.git cd Ollama-mcp-client pip install -r requirements.txt ``` ### Configuration 1. Configure my MCP server in `client/mcp_server_config.json`. My mcp server is released in the release, with the following format: ```json { "mcpServers": { "server-name": { "url": "http://localhost:8080/sse" } } } ``` If you have your own mcp-server, you can also configure your own server address. 2. Configure the Ollama model in `client/Constants.py`: ```python LLAMA_MODEL_QWEN = "qwen3:14b" # Or other Ollama-supported models ``` ### Run ```bash python client/main.py ``` ## Project Structure ``` Ollama-MCP-Client ├── client │   ├── Constants.py │   ├── McpClient.py │   ├── OllamaAgent.py │   ├── OllamaTools.py │   ├── __init__.py │   ├── main.py │   ├── mcp_server_config.json │   └── utils │   ├── JsonUtil.py ├── requirements.txt └── test ├── __init__.py └── test_ollamaToolsformat.py ``` ## Acknowledgements This project was inspired by [mihirrd/ollama-mcp-client](https://github.com/mihirrd/ollama-mcp-client). Their project only supports local and does not support sse, while mine is the opposite. Thank you for that. ## License `MIT License`