Skip to content

Latest commit

 

History

History
38 lines (23 loc) · 897 Bytes

README.md

File metadata and controls

38 lines (23 loc) · 897 Bytes

Model Context Protocol

This example demonstrates how to implement an agent that can use tools provided by MCP (Model Context Protocol) servers.

Prerequisites

Quick Start

First start the MCP server in one terminal:

python examples/mcp/server.py

Then start the MCP agent in another terminal:

export MODEL_ID="your-model-id"
export MODEL_BASE_URL="your-base-url"
export MODEL_API_VERSION="your-api-version"
export MODEL_API_KEY="your-api-key"

python examples/mcp/agent.py

Finally, communicate with the agent using the coagent CLI:

coagent mcp -H type:ChatMessage -F .content.content -d '{"role":"user","content":"What is the weather like in Beijing"}'