MCP Ollama Agent

🚀 TypeScript MCP Agent Integration with Ollama

This project demonstrates the integration between a Model Context Protocol (MCP) server and Ollama, allowing AI models to interact with various tools through a unified interface.

✨ Features

• Supports multiple MCP servers (tested with uvx and npx).
• Built - in file system operations and web search capabilities.
• Configurable via mcp-config.json, similar to claude_desktop_config.json.
• Features an interactive chat interface, supporting integration with Ollama and the use of any tool.
• An independent demo mode for testing web and file system tools without an LLM.

🚀 Quick Start

Prerequisites

• Node.js (version 18 or higher).
• Ollama installed and running.
• Install the required MCP tools globally:

# File system operations
npm install -g @modelcontextprotocol/server-filesystem

# Web search
npm install -g @mzxrai/mcp-webresearch

Clone and Install

git clone https://github.com/ausboss/mcp-ollama-agent.git
cd mcp-ollama-agent
npm install

Configure Tools and Ollama Model in mcp-config.json

{
    "mcpServers": {
        "filesystem": {
            "command": "npx",
            "args": ["@modelcontextprotocol/server-filesystem", "./"]
        },
        "webresearch": {
            "command": "npx",
            "args": ["-y", "@mzxrai/mcp-webresearch"]
        }
    },
    "ollama": {
        "host": "http://localhost:11434",
        "model": "qwen2.5:latest"
    }
}

Run the Demo Mode to Test File System and Web Search Tools without an LLM

npx tsx ./src/demo.ts

Or Start the Chat Interface with Ollama

npm start

🔧 Configuration

• MCP Servers: Add any MCP - compatible server to the mcpServers section.
• Ollama: Configure the host and model (must support function calls).
• Supports MCP servers for Python (uvx) and Node.js (npx).

💻 Usage Examples

The model used is qwen2.5:latest

Chat started. Type "exit" to end the conversation.
You: Can you use your list directory tool to view the contents in test - directory and then use the read file tool to send its contents to me?
The model is using tools to assist in answering...
Tool used: list_directory
Parameters: { path: 'test-directory' }
Tool result: [ { type: 'text', text: '[FILE] test.txt' } ]
Assistant:
The model is using tools to assist in answering...
Tool used: read_file
Parameters: { path: 'test-directory/test.txt' }
Tool result: [ { type: 'text', text: 'rosebud' } ]
Assistant: The content of the `test.txt` file in `test-directory` is:
rosebud
You: Thank you
Assistant: You're welcome! If you have other needs or require further assistance, please feel free to let me know.

🤝 Contribution

Contributions are welcome! Feel free to submit issues or pull requests.