Agentic Researcher

A fully autonomous, verifiable research agent powered by LangGraph, Mistral AI, and Chainlit.

Agentic Researcher doesn't just "chat"—it plans, searches the live web, scrapes valid sources, builds a local knowledge base, writes comprehensive technical reports, and self-corrects its own work before presenting it to you.

---

🚀 Key Features

• 🧠 Autonomous Planning: Breaks down complex user queries into specific, researchable sub-topics.
• 🔄 Self-Correcting Loop: Includes a "Review Node" that evaluates the quality of the report. If the data is insufficient, it autonomously loops back to do more research.
• 🕸️ Interactive Knowledge Graphs: Automatically extracts entities (technologies, organizations, concepts) and generates an interactive HTML network graph to visualize connections.
• 📄 PDF Report Generation: Produces professional PDF reports with citations and metadata.
• 📚 RAG-based Q&A: After research is complete, the agent switches to "Expert Chat Mode," allowing you to ask follow-up questions based strictly on the collected data.
• 🛡️ Quality Filters: Automatically filters out low-quality sources (social media, spam) and uses fallback mechanisms for scraping failures.

🛠️ Technical Architecture

The system is built on a cyclic graph architecture using LangGraph:

graph TD
    A[Start] --> B[Plan Node]
    B --> C[Research Node]
    C --> D[Scrape Node]
    D --> E[Write Node]
    E --> F[Review Node]
    F -->|Needs More Info| C
    F -->|Report Complete| G[End]

Loading

Project Structure

├── app.py              # Main Chainlit application entry point
├── agent_graph/        # Core logic
│   ├── graph.py        # LangGraph definition & workflow
│   ├── nodes.py        # The specific functions (Plan, Research, Write, etc.)
│   ├── states.py       # Pydantic/TypedDict state definitions
│   └── global_state.py # Singleton for ChromaDB persistence
├── utils/
│   ├── pdf_gen.py      # PDF generation logic (xhtml2pdf)
│   └── graph_viz.py    # Knowledge graph generation (NetworkX/Pyvis)
├── tools/
│   ├── search.py       # Serper API wrapper
│   └── scraper.py      # Browserless API wrapper
└── requirements.txt    # Project dependencies

⚙️ Setup Instructions

Follow these steps to set up the project locally.

1. Clone the Repository

git clone https://github.com/nitesh-77/Agentic-Researcher.git

2. Create a Virtual Environment

# Create a new conda environment 
conda create -n research-agent python=3.10 -y

# Activate the environment
conda activate research-agent

3. Install the required packages

pip install -r requirements.txt

4. Configure API keys

Rename the environment file: Change the template file example.env to .env

Add your API Keys:

MISTRAL_API_KEY=your_mistral_api_key_here
SERP_API_KEY=your_serper_dev_key_here
BROWSERLESS_API_KEY=your_browserless_io_key_here

5. Run the application

chainlit run app/app.py