LLM failed after calling a tool from a user-defined MCP server

[sebastiengilbert73]

The symptom is identical to this other topic (for which there are no answers yet):

Anyone managed to use mcp server with ollama model provider?

Hey everyone , I am trying to use a local llm while using mcp servers in crew ai i can see in verbose log that llm is able to identify correct tool and able to run it and receive output from tool but after that step it throws an APIConnectionError Here is my code snippet from crewai import Agent, Task, Crew, LLM, Process from crewai_tools import MCPServerAdapter from mcp import StdioServerParameters import os from crewai import LLM server_params=StdioServerParameters( command="uv", # O…