MCP Hub is a research and code assistant built using Gradio's MCP server functionality, demonstrating how to build a workflow of connected AI agents that work together. The system orchestrates an 8-step process, including question enhancement, web search, LLM summarization, citation formatting, result combination, code generation, code execution in a sandbox environment, and final summary creation. It integrates with services like Tavily API and Nebius LLMs to provide up-to-date information and powerful text processing capabilities, ultimately delivering executable Python code based on research findings.
Key Features
01Real-time Web Search via Tavily API integration
02Sandboxed Code Execution using Modal
03Multi-Agent Architecture for interconnected agent services
040 GitHub stars
05Automated APA-style citation formatting
06Code Generation of executable Python code based on research
