McpLLMServer

import httpx OLLAMA_URL = "http://ollama:11434/api/generate" MODEL = "llama3" async def generate(prompt: str) -> str: async with httpx.AsyncClient(timeout=60) as client: response = await client.post( OLLAMA_URL, json={ "model": MODEL, "prompt": prompt, "stream": False } ) response.raise_for_status() return response.json()["response"]