MCP Rubber Duck

Example Use Cases

• 1

  Compare and aggregate responses from multiple OpenAI-compatible LLM providers to get diverse perspectives or perform model benchmarking.

• 2

  Persistent conversational context across messages and sessions, including conversation clearing and model-specific conversation switching.

• 3

  Access external MCP servers/tools (documentation, file access, APIs) through an MCP bridge with configurable trust and session-based approvals to minimize token usage.

Description

MCP Rubber Duck is an open-source MCP server that bridges multiple OpenAI-compatible LLM endpoints (OpenAI, Google Gemini, Anthropic via compatible endpoints, Groq, Ollama, and more). It lets you query multiple providers simultaneously, maintain conversation context, cache responses, and automatically failover between providers. The server can act as an MCP server to clients (e.g., Claude Desktop) and as an MCP bridge to access tools on external MCP servers with session-based approval controls and per-server trust configuration.