A TypeScript SDK for building agent-native apps with generative UI, shared state, and human-in-the-loop flows in React, Angular, Vue, and chat apps.
CopilotKit/open-research-ANA
A demo research canvas app where an AI agent runs real-time web search with human-in-the-loop review, built on CopilotKit, Tavily and LangGraph.
- Stars
- 423
- Forks
- 113
- Open issues
- 3
- Last commit
- 2 days ago
What is it?
open-research-ANA is a demo of an Agent Native Application: a research canvas where a user and an AI agent work on the same document. The agent runs real-time searches through Tavily and pauses for human input at key steps, so the user can steer the research rather than only reading a final answer.
The agent side is built with LangGraph, and the frontend uses CopilotKit to connect the chat interface and canvas to the running agent. Running it locally involves starting the agent with the LangGraph CLI, opening a tunnel with the CopilotKit CLI, and launching a Next.js frontend with pnpm. It requires OpenAI, Tavily, LangSmith and CopilotKit API keys.
The repository has been consolidated into the CopilotKit monorepo, where the maintained version lives under examples/showcases/research-canvas.
Key features
- 01Real-time web search powered by Tavily
- 02Human-in-the-loop checkpoints where the user reviews agent steps
- 03Agent orchestration built with LangGraph
- 04Chat and canvas frontend connected to the agent via CopilotKit
- 05Local setup with the LangGraph CLI and a Next.js frontend
Use cases
- Exploring a research topic with an agent that searches the web and asks for guidance
- Learning how to combine LangGraph agents with a CopilotKit frontend
- Prototyping a human-in-the-loop research assistant
Quick start
- 1
Install the prerequisites: pnpm, Docker and the LangGraph CLI
- 2
Start the agent from the agent folder
langgraph up - 3
Install the frontend dependencies from the frontend folder
pnpm install - 4
Start the frontend dev server
pnpm run dev
Things to know
- Running it requires accounts and API keys for OpenAI, Tavily, LangSmith and CopilotKit.
- The project is a demo and has been consolidated into the CopilotKit monorepo, where the latest version lives.
Topics
FAQ
- What is open-research-ANA?
- A demo research canvas app where an AI agent runs real-time web search with human-in-the-loop review, built on CopilotKit, Tavily and LangGraph. open-research-ANA is a demo of an Agent Native Application: a research canvas where a user and an AI agent work on the same document. The agent runs real-time searches through Tavily and pauses for human input at key steps, so the user can steer the research rather than only reading a final answer.
- Is open-research-ANA free to use? What license does it have?
- open-research-ANA has no declared license. Its code is public, but without a license you have no explicit permission to reuse it; ask the author before relying on it.
- Is open-research-ANA still maintained?
- Actively maintained. The last commit to open-research-ANA was on Oct 19, 2026.
- What can I use open-research-ANA for?
- Exploring a research topic with an agent that searches the web and asks for guidance Learning how to combine LangGraph agents with a CopilotKit frontend Prototyping a human-in-the-loop research assistant
- How do I install open-research-ANA?
- Install the prerequisites: pnpm, Docker and the LangGraph CLI Start the agent from the agent folder: langgraph up Install the frontend dependencies from the frontend folder: pnpm install Start the frontend dev server: pnpm run dev
- Where can I get open-research-ANA?
- The source code and installation instructions are on GitHub: https://github.com/CopilotKit/open-research-ANA
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
An opinionated, batteries-included agent harness for Python that runs long multi-step tasks out of the box on LangGraph.
An LLM agent framework from Microsoft that automates data-driven R&D tasks such as reading papers, implementing models and factors, and running Kaggle-style experiments.
Python SDK for building agents on Claude Code, with query(), an interactive client, in-process custom tools and hooks.
A VS Code extension that turns plain .chat.md files into editable AI chat sessions with MCP tool support across LLM providers.
Agent-native memory infrastructure that turns LLM agent conversations and execution into persistent, structured memory for Python and TypeScript apps.