andybrandt/mcp-simple-openai-assistant

Overview

A Python-based MCP server that enables Claude to call and interact with OpenAI GPT assistants. It routes requests from Claude's tool-use interface through the Model Context Protocol to OpenAI's API, letting developers chain models across providers.

Best for

Best for
Developers experimenting with multi-model agent workflows

Use cases

• Bridge Claude to OpenAI's assistant API for hybrid workflows
• Give Claude access to custom GPTs with specialized knowledge
• Build multi-model pipelines without manual API switching

How to use

Install

pip install mcp-simple-openai-assistant

Tools exposed

• create_assistant
• list_assistants
• retrieve_assistant
• update_assistant
• create_new_assistant_thread
• list_threads
• delete_thread
• ask_assistant_in_thread

Tested with

Claude Desktop

Example client config

{\n  "mcpServers": {\n    "openai-assistant": {\n      "command": "python",\n      "args": ["-m", "mcp_simple_openai_assistant"],\n      "env": {\n        "OPENAI_API_KEY": "your-api-key-here"\n      }\n    }\n  }\n}

Notes

A Python-based MCP server that enables Claude to call and interact with OpenAI GPT assistants. It routes requests from Claude’s tool-use interface through the Model Context Protocol to OpenAI’s API, letting developers chain models across providers.

37 stars on GitHub. Last updated 2025-09-12. Licensed MIT.

Use cases

• Bridge Claude to OpenAI’s assistant API for hybrid workflows
• Give Claude access to custom GPTs with specialized knowledge
• Build multi-model pipelines without manual API switching

Pros

• Enables cross-provider model orchestration with minimal code
• Lightweight Python server, easy to self-host
• Leverages existing OpenAI assistant configurations

Cons

• Requires API keys for both Claude and OpenAI, increasing cost
• Limited to 37 stars, small community for support
• No streaming or real-time response handling documented

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.