
Rubber Duck
Wraps any OpenAI-compatible API endpoint plus CLI coding agents like Claude Code, Aider, and Gemini CLI into MCP tools for querying multiple LLMs simultaneously. Exposes operations for single queries, multi-model comparisons, structured debates between models, consensus voting, and iterative response refinement. Includes conversation management, automatic failover, usage tracking, and an MCP bridge that lets your "ducks" access other MCP servers. You'd reach for this when debugging complex problems that benefit from multiple AI perspectives, comparing model outputs side-by-side, or running structured evaluations where models judge each other's responses. Supports rich HTML interfaces in MCP Apps-compatible clients.
- GitHub stars
- 167
- tools
- 15
- auth
- Secret needed
- transport
- STDIO
- env vars
- 4
- last updated
- Apr 3, 2026
Tools
15 toolsPublic tool metadata for what this MCP can expose to an agent.
ask_duckAsk a single question to a specific LLM provider
Ask a single question to a specific LLM provider
No parameter schema in public metadata yet.
chat_with_duckConversation with context maintained across messages
Conversation with context maintained across messages
No parameter schema in public metadata yet.
clear_conversationsClear all conversation history
Clear all conversation history
No parameter schema in public metadata yet.
list_ducksList configured providers and health status
List configured providers and health status
No parameter schema in public metadata yet.
list_modelsList available models for providers
List available models for providers
No parameter schema in public metadata yet.
compare_ducksAsk the same question to multiple providers simultaneously
Ask the same question to multiple providers simultaneously
No parameter schema in public metadata yet.
duck_councilGet responses from all configured ducks
Get responses from all configured ducks
No parameter schema in public metadata yet.
get_usage_statsUsage statistics and estimated costs
Usage statistics and estimated costs
No parameter schema in public metadata yet.
duck_voteMulti-duck voting with reasoning and confidence
Multi-duck voting with reasoning and confidence
No parameter schema in public metadata yet.
duck_judgeHave one duck evaluate and rank others' responses
Have one duck evaluate and rank others' responses
No parameter schema in public metadata yet.
duck_iterateIteratively refine a response between two ducks
Iteratively refine a response between two ducks
No parameter schema in public metadata yet.
duck_debateStructured multi-round debate between ducks
Structured multi-round debate between ducks
No parameter schema in public metadata yet.
mcp_statusMCP Bridge status and connected servers
MCP Bridge status and connected servers
No parameter schema in public metadata yet.
get_pending_approvalsPending MCP tool approval requests
Pending MCP tool approval requests
No parameter schema in public metadata yet.
approve_mcp_requestApprove or deny a duck's MCP tool request
Approve or deny a duck's MCP tool request
No parameter schema in public metadata yet.
Sponsors
Configuration
OPENAI_API_KEYsecretOpenAI API key (starts with sk-)
GEMINI_API_KEYsecretGoogle Gemini API key
GROQ_API_KEYsecretGroq API key (starts with gsk_)
DEFAULT_PROVIDERDefault LLM provider to use
Readme
MCP Rubber Duck
An MCP (Model Context Protocol) server that acts as a bridge to query multiple LLMs -- both OpenAI-compatible HTTP APIs and CLI coding agents. Just like rubber duck debugging, explain your problems to various AI "ducks" and get different perspectives!
MCP Rubber Duck - AI ducks helping debug code
Details
- Repository
- github.com/nesquikm/mcp-rubber-duck
- Package
mcp-rubber-duck- Transport
- STDIO
- Auth
- Required
- Registry
- MCP registry, active