ResearchMind — Multi-Agent Research Pipeline

🚀 Live Demo: https://multi-agent-research-systems.streamlit.app/

An AI-powered research system where four specialized agents collaborate to search the web, scrape content, write a structured report, and critically review it — all from a single topic input.

Built with LangChain, Mistral AI, Tavily Search, and Streamlit.

---

How It Works

The pipeline runs in 4 sequential steps:

Step	Agent	Role
01	Search Agent	Searches the web via Tavily and collects titles, URLs, and snippets
02	Reader Agent	Picks the most relevant URL and scrapes its full content
03	Writer Chain	Combines research and writes a structured report using an LLM
04	Critic Chain	Reviews the report and gives a score, strengths, and areas to improve

---

Tech Stack

• LangChain — agent orchestration and chain composition
• Mistral AI — LLM backend for writing and critiquing
• Tavily API — real-time web search
• BeautifulSoup — web scraping and content extraction
• Streamlit — frontend UI
• python-dotenv — environment variable management

---

Project Structure

├── app.py              # Streamlit UI
├── agents.py           # Search agent, Reader agent, Writer chain, Critic chain
├── tools.py            # web_search and scrape_url tools
├── pipeline.py         # Core pipeline logic (CLI entry point)
├── requirements.txt    # Dependencies
└── .env                # API keys (never commit this)

---

Getting Started

1. Clone the repo

git clone https://github.com/tahirshamim/multi-agent-research-system.git
cd multi-agent-research-system

2. Create a virtual environment

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

3. Install dependencies

pip install -r requirements.txt

4. Set up environment variables

Create a .env file in the project root:

MISTRAL_API_KEY=your-mistral-key-here
TAVILY_API_KEY=your-tavily-key-here

Get your keys:

• Mistral → console.mistral.ai
• Tavily → app.tavily.com

5. Run the app

# Streamlit UI
streamlit run app.py

# Or CLI mode
python pipeline.py

---

Deployment

Deployed on Streamlit Cloud. To deploy your own:

1. Push the repo to GitHub (without .env)
2. Go to share.streamlit.io
3. Connect your repo and set main file as app.py
4. Add your API keys under Settings → Secrets:

MISTRAL_API_KEY = "your-mistral-key-here"
TAVILY_API_KEY = "your-tavily-key-here"

---

Author

Tahir Shamim
CS Student @ NED University Karachi
github.com/tahirshamim