Open Research Ana
🤖 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
Last verified:
What is Open Research Ana?
Open Research Ana is an open-source, AI agent-native research canvas application that performs real-time web search with Human-in-the-Loop (HITL) capabilities. Powered by CopilotKit, Tavily, and LangGraph, it simplifies complex research tasks by combining human creativity with agent efficiency to make research projects faster and smarter.
Key features include real-time web retrieval powered by Tavily for accurate, up-to-date information; Human-in-the-Loop UX that encourages step-by-step collaboration between humans and agents; agent-native workflow built on LangChain's LangGraph and CopilotKit's Co-Agents; and generative UI that streams insights as they're found rather than just at the end. The application allows users to ask research questions in chat, propose research scopes, pause at key points to focus research, and receive explainable insight cards.
This tool is designed for researchers, students, academics, developers, and professionals who need to conduct web-based research, write reports, synthesize information from multiple sources, and collaborate with AI agents. It's particularly valuable for those who want to customize and adapt their research workflows, build their own AI research applications, or explore human-agent collaboration patterns.
The application streams the agent's actions in real-time so users aren't bored waiting, extracts structured data for insight pills, and supports research across news, patents, papers, and other web content. Users can set time windows for research (last week, last month, last year) and pick what innovation areas matter most to them.
Open Research Ana pricing
Pricing model: Freemium
Free and open source. The application is completely free to use as it is an open-source project hosted on GitHub. However, users running it locally need to pay for their own API keys: OpenAI API, Tavily API, LangSmith API, and CopilotKit API, which have their own separate pricing structures.
Open Research Ana pros
- Open source for exploration and customization
- Real-time web retrieval with Tavily integration
- Human-in-the-loop UX for iterative research
- Agent-native workflow with LangGraph
- Generative UI streams insights as found
- Explains research process with actionable cards
- Combines human creativity with agent efficiency
- Supports multiple research scopes simultaneously
- Real-time streaming of agent actions
- Structured output extraction for insights
- Works with news, patents, and academic papers
- Time window filtering for research (week/month/year)
- CopilotKit Co-Agents infrastructure
- Interactive and efficient research tasks
- Builds trust through transparent process
Open Research Ana cons
- Requires multiple API keys (OpenAI, Tavily, LangSmith, CopilotKit)
- Needs local setup with pnpm, Docker, and Langgraph CLI
- No pre-built releases available yet
- Requires creating tunnel to local agent
- Technical setup complexity for non-developers
- No official documentation beyond CopilotKit docs
- Still in demo/showcase stage
- Dependent on external API costs
Frequently asked questions about Open Research Ana
What is Open Research Ana?
Open Research Ana is an open-source, AI agent-native research canvas application that performs real-time search with Human-in-the-Loop (HITL) capabilities. It combines human creativity with agent efficiency using CopilotKit, Tavily, and LangGraph to make research projects faster and smarter.
Is Open Research Ana free to use?
Yes, Open Research Ana is completely free and open source. However, running it locally requires your own API keys for OpenAI, Tavily, LangSmith, and CopilotKit, which have their own separate pricing.
What API keys do I need to run it locally?
You need four API keys: OpenAI API key, Tavily API key, LangSmith API key, and CopilotKit API key. These are required to create the .env files for both the agent and frontend components.
What technologies power Open Research Ana?
It is powered by LangGraph for agent architecture, Tavily for real-time web retrieval, CopilotKit for the agentic interface and Co-Agents, and LangChain's LangGraph for human-in-the-loop workflows.
What is Human-in-the-Loop (HITL) in this context?
HITL means the agent pauses at key points during research so you can focus the research, pick what innovation areas matter to you, set time windows, add details, and approve the research scope before the agent continues searching.
What types of content can it search?
Open Research Ana can search across news, patents, academic papers (including arXiv), and general web content through Tavily's real-time web retrieval capabilities.
How do I start the application?
First start the agent by running 'langgraph up' in the agent directory after setting up your .env file. Then create a tunnel with 'npx copilotkit@latest dev --port 8123'. Finally, install dependencies and run 'pnpm run dev' in the frontend directory.
What programming languages is it built with?
The codebase is 68.2% TypeScript, 29.6% Python, 2.1% CSS, and 0.1% JavaScript. The frontend is TypeScript-based while the agent uses Python.
Can I customize the research workflow?
Yes, being open source, it is designed for exploration, customization, and adaptation to suit your unique workflows. You can modify the agent behavior, UI components, and integration points.
What makes it different from regular AI chatbots?
Unlike regular chatbots, it has an agent-native workflow with Human-in-the-Loop UX, streams insights as they're found with generative UI, provides explainable research cards, and allows you to pause and direct the research process at key points rather than just getting a final answer.