Hydra

关于

HydraMCP serves as an MCP server designed to empower AI agents within environments like Claude Code, allowing them to interact with and manage various large language models (LLMs) through existing subscriptions. It eliminates the need for extra API keys or per-token billing, integrating cloud services like OpenAI, Google Gemini, and Anthropic Claude with local models via Ollama. This tool streamlines the process of querying, comparing, and synthesizing responses from diverse LLMs, providing powerful multi-model orchestration capabilities directly within your development workflow.

主要功能

• Query any available LLM and get a response back
• Compare up to 5 models in parallel with a single prompt
• Achieve consensus among multiple models using an LLM-as-judge approach
• Synthesize superior answers by combining the best ideas from various models
• List all available models across different providers
• 1 GitHub stars

使用案例

• Comparing different LLMs (e.g., Gemini and Claude) on a technical approach
• Reviewing code functions using a specific large language model like GPT-5
• Getting consensus from multiple models on code safety or design principles