Atlas

A multi-agent research pipeline powered by Google Gemini. Given a topic, Atlas runs it through three specialized agents — a researcher, a critic, and a summarizer — and returns a clean, validated summary via a FastAPI gateway.

---

How it works

Topic → [ Research Agent ] → [ Critic Agent ] → [ Summarizer Agent ] → Summary

1. Research Agent — gathers findings on the given topic
2. Critic Agent — validates the findings, filtering out anything unsupported
3. Summarizer Agent — produces a structured summary from the verified findings only

---

Tech stack

• Backend — FastAPI
• Agents — pydantic-ai
• Model — Google Gemini (via google-generativeai)
• Validation — Pydantic v2

---

Prerequisites

• Python 3.11+
• A Google Gemini API key

---

Getting started

1. Clone the repo

git clone https://github.com/mgalen007/atlas.git
cd atlas

2. Create and activate a virtual environment

python -m venv .venv
source .venv/bin/activate  # Windows: .venv\Scripts\activate

3. Install dependencies

pip install -r requirements.txt

4. Set up environment variables

cp .env.example .env

Open .env and add your Gemini API key:

GEMINI_API_KEY=your_key_here

5. Run the server

cd server
uvicorn main:app --reload

The API will be available at http://localhost:8000.

---

API

GET /api/research?topic={topic}

Runs the full research pipeline on a given topic.

Example request

curl "http://localhost:8000/api/research?topic=quantum+computing"

Example response

{
  "success": true,
  "data": {
    "topic": "Quantum computing",
    "key_findings": [
      "Quantum computers use qubits instead of classical bits.",
      "Superposition allows qubits to represent multiple states simultaneously."
    ],
    "content": "Quantum computing represents a fundamental shift in how computation..."
  }
}

GET /api/health-check

A health check endpoint, use it to verify if the server is running correctly.

GET /docs

The Swagger UI documentation for the API.

---

Project structure

atlas/
├── server/
│   ├── main.py
│   └── features/
│       ├── __init__.py
│       ├── research/
│       ├── __init__.py
│       │   └── router.py
│       └── agents/
│           ├── __init__.py 
│           ├── config/
│           │   └── models.py
│           ├── service.py
│           ├── research_agent.py
│           ├── critic_agent.py
│           └── summarizer_agent.py
├── client/               # frontend (coming soon)
├── .env.example
├── requirements.txt
└── README.md

---

Screenshots

Swagger UI docs

Example response in Postman for topic "Generative AI"

---

License

MIT