open-research-ANA
π€ An open-source, AI agent-native research canvas application that performs real-time search with HITL (Human in The Loop) capabilities, powered by CopilotKit,
What is open-research-ANA?
π€ An open-source, AI agent-native research canvas application that performs real-time search with HITL (Human in The Loop) capabilities, powered by CopilotKit, Tavily and LangGraph
Overview of open-research-ANA
What it is
open-research-ANA is an open-source demo application that combines a research canvas with an AI agent-native interface. It uses Tavily for real-time web search and CopilotKit for agentic interactions, with Human-in-the-Loop (HITL) capabilities. The app is powered by LangGraph. It is part of the CopilotKit monorepo.
Key features
- Real-time search via Tavily - Human-in-the-Loop (HITL) capabilities for interactive research - Agent-native interface built with CopilotKit - Powered by LangGraph for complex research tasks
Use cases
- Conducting interactive research where a user can guide an AI agent in real time - Demonstrating Human-in-the-Loop workflows in an AI application - Exploring how to combine search, agents, and user feedback in a single canvas app
Getting started
See the repository README for detailed setup instructions. The quick start requires pnpm, Docker, and Langgraph CLI. API keys for OpenAI, Tavily, LangSmith, and CopilotKit are needed. To start the agent, go to the agent directory, set up a .env file with the keys, and run langgraph up. Open a tunnel to the local agent with npx copilotkit@latest dev --port 8123. Then, in the frontend directory, run pnpm install, create a .env file with the frontend keys, and start the dev server with pnpm run dev.
License and pricing
The license is not stated in the provided metadata or README. The project is listed with pricing_type as open_source.
Based on the project's public README and repository metadata.