♾️ Intellex

Intellex is a multi-agent AI research assistant that turns complex topics into clear, consolidated research reports.

It automates deep research with artificial analysts and experts. It simulates a team of domain analysts conducting interviews with virtual domain experts to produce a consolidated Markdown report on the given topic.

♾️ Features

• Parallel virtual analysts simulated with Anthropic models
• Expert interviews orchestration powered by LangGraph
• Markdown-ready final reports
• Want to try? just Plug-and-play

♾️ WorkFlow

1. Provide a research topic + number of analysts.
2. Intellex spawns analysts who each interview an expert.
3. Each analyst produces a mini-report.
4. Reports are merged into one consolidated Markdown file.

♾️ Project Architecture

• Using LangGraph to orchestrate the team of analysts and their interviews.
• The core logic follows a Map-Reduce pattern, where analysts independently interview the experts in parallel in the "map" phase, then synthesize their individual findings into a single, comprehensive report in the "reduce" phase.
• The entire process is powered by Anthropic's language models.

♾️ Quickstart

Prerequisites

• Python 3.11+ (python 3.13 recommended)
• uv installed (pip install uv)

Follow these simple steps to get the project running locally.

1. Clone the repository

git clone https://github.com/your-username/research-assistant.git
cd research-assistant

2. Sync the project with original project dependencies and versions

# just run this simple command and you've done!
uv sync

3. Create a .env file from env.example in the project root and add your API keys.

4. Start the LangGraph studio to interact with the agent.

cd agent/
langgraph dev

5. Once the server is running, the LangGraph Studio WebUI will open in the browser, and you can interact with it.

🤝 Contributing

Feedback and PRs are welcome! Please open an issue before major changes. I am still working on it.

Happy Researching! 🔍