This project provides an AI agent-native research canvas application, ANA, designed for interactive and efficient research. It targets researchers and power users by integrating real-time search with human-in-the-loop capabilities, powered by CopilotKit, Tavily, and LangGraph.

How It Works

ANA leverages LangGraph to orchestrate a research workflow. It combines Tavily for real-time web search with CopilotKit for building an agentic interface, allowing for human intervention and guidance during the research process. This approach aims to enhance the efficiency and interactivity of complex research tasks.

Quick Start & Requirements

• Install/Run: Requires pnpm, Docker, and LangGraph CLI.
• API Keys: Needs OpenAI, Tavily, LangSmith, and CopilotKit API keys.
• Setup: Involves starting the agent (cd agent, create .env, langgraph up) and then the frontend (cd frontend, pnpm install, create .env, pnpm run dev). A tunnel to the local agent is also required (npx copilotkit@latest dev --port 8123).
• Documentation: CopilotKit Docs, LangGraph Platform Docs

Highlighted Details

• Agent-native research canvas application.
• Integrates real-time search (Tavily) with Human-in-the-Loop (HITL).
• Powered by CopilotKit for agentic interface and LangGraph for orchestration.

Maintenance & Community

No specific details on contributors, sponsorships, or community channels are provided in the README.

Licensing & Compatibility

The README does not specify a license. Compatibility for commercial use or closed-source linking is not mentioned.

Limitations & Caveats

The project requires multiple API keys and specific tooling (pnpm, Docker, LangGraph CLI), which may increase the barrier to entry. The README does not detail the project's license or commercial use implications.