Agent Integration

Claude Desktop Integration

Configure Claude Desktop to use the Functor MCP Server by updating your claude_desktop_config.json.

{
  "mcpServers": {
    "functor": {
      "command": "python",
      "args": ["app/mcp/server.py"],
      "env": {
        "MCP_API_KEY": "your-api-key",
        "QDRANT_URL": "http://localhost:6333",
        "REDIS_URL": "redis://localhost:6379",
        "LLM_PROVIDER": "gemini"
      }
    }
  }
}

Multi-Server Integration

The Functor MCP server can work alongside other MCP servers to provide a comprehensive knowledge ecosystem.

{
  "mcpServers": {
    "functor": {
      "command": "python",
      "args": ["app/mcp/server.py"],
      "env": { "MCP_API_KEY": "..." }
    },
    "database": {
      "command": "python",
      "args": ["database_mcp_server.py"],
      "env": { "DB_CONNECTION_STRING": "..." }
    },
    "web-search": {
      "command": "python",
      "args": ["web_search_mcp.py"],
      "env": { "SEARCH_API_KEY": "..." }
    }
  }
}

Python Agent Integration

Integrate with custom Python agents using the FunctorClient.

from functor_mcp_sdk import FunctorClient

class AIAssistant:
    def __init__(self):
        self.client = create_client(api_key="your-api-key")

    async def process_query(self, query: str):
        async with self.client.session():
            # Store query in episodic memory
            await self.client.episodic_add_episode(
                event_type="query",
                event_data={"query": query}
            )

            # Process query with smart routing
            result = await self.client.execute_smart(
                query=query,
                enable_analytics=True
            )

            return result