MAD (Multi-Agents-Debate) is an open-source framework exploring the collective debating capabilities of LLMs. It addresses the 'Degeneration of Thoughts' (DoT) issue—where self-reflection leads to biased or rigid outputs—by implementing a tit-for-tat interaction between multiple agents. By assigning roles like 'devil' and 'angel' to provide mutual external feedback, MAD corrects distorted thinking and breaks cognitive rigidity. It achieves significant performance gains in Counterintuitive QA and Commonsense Machine Translation tasks.
mcp-use is a fullstack MCP framework designed to build interactive MCP Apps and Servers for AI Agents like ChatGPT and Claude. It provides TypeScript and Python SDKs, enabling developers to create 'write once, run everywhere' UI widgets. The ecosystem includes an integrated MCP Inspector for debugging and Manufact MCP Cloud for production deployment with full observability, metrics, and logs, simplifying the creation of interactive AI components.
awesome-llm-apps by Shubhamsaboo is a curated repository offering over 100 runnable AI Agent and RAG application templates. It serves as a practical cookbook of ready-to-use code, enabling developers to quickly clone, customize, and deploy production-grade LLM applications. It covers modern AI stacks like AI Agents, multi-agent teams, RAG, voice agents, agent skills, and fine-tuning. Each template is self-contained, original, end-to-end tested, supports various LLMs (Claude, Gemini, OpenAI, Llama, etc.), and comes with free step-by-step tutorials.