MCPChatbot Example

What is MCPChatbot Example?

MCPChatbot Example is a demonstration project integrating the Model Context Protocol (MCP) with customized LLMs (e.g., Qwen) to create a chatbot that can interact with external tools through MCP servers. It runs on Python 3.10+ and includes CLI, terminal, and Streamlit web interfaces for developers exploring MCP-based tool‑augmented LLM applications.

How to use MCPChatbot Example?

Clone the repository, set up a Python virtual environment, install dependencies via pip install -r requirements.txt, configure the .env file with LLM API keys and folder paths, then modify mcp_servers/servers_config.json to point to local paths. Run examples from the example/ directory, e.g., streamlit run example/chatbot_streamlit/app.py for the web chatbot.

Key features of MCPChatbot Example

• Supports multiple LLM providers (e.g., Qwen, Ollama)
• Includes CLI, terminal, and Streamlit web chatbot interfaces
• Provides both regular and streaming response modes
• Integrates built‑in MCP servers (e.g., Markdown processing tools)
• Offers single‑prompt examples and interactive chat examples
• Visualizes MCP tool workflow in the Streamlit sidebar

Use cases of MCPChatbot Example

• Prototyping an MCP‑enabled chatbot with custom LLM backends
• Demonstrating LLM tool calls through MCP servers (e.g., Markdown processing)
• Testing single‑prompt interactions in regular or streaming mode
• Building an interactive terminal chatbot that chains multiple MCP tool calls
• Creating a web‑based chatbot with real‑time streaming and tool workflow traces

FAQ from MCPChatbot Example

What LLM models does MCPChatbot Example support?

It supports customized LLMs via the OpenAI‑compatible API (e.g., Qwen) and Ollama, configured through the .env file.

How do I configure MCP servers?

Edit mcp_servers/servers_config.json with the absolute path to your uv executable and the project’s mcp_servers directory. Windows users should adjust the command and path accordingly.

What are the software requirements?

Python 3.10 or higher, and dependencies including python-dotenv, mcp[cli], openai, and colorama (automatically installed).

Is there a web interface?

Yes, a Streamlit web chatbot example is included; run streamlit run example/chatbot_streamlit/app.py.

Can I extend the project with new MCP servers?

Yes, you can add new MCP servers in mcp_servers/, update servers_config.json, and implement new functionalities.