Deep Research

An AI-powered deep research system built with LangChain and LangGraph. It autonomously plans research, searches the web, gathers information, and generates comprehensive reports.

Architecture

┌────────────────────────────────────────────────────┐
│                 LangGraph Workflow                   │
│                                                     │
│  ┌──────────┐    ┌────────────────┐   ┌──────────┐ │
│  │ Planner  │───▶│ Research Agent │──▶│  Report  │ │
│  │ (Chain)  │    │    (ReAct)     │   │Generator │ │
│  └──────────┘    └───────┬────────┘   │ (Chain)  │ │
│                          │            └──────────┘ │
│                 ┌────────┼────────┐                 │
│                 ▼        ▼        ▼                 │
│            web_search  get_urls  fetch_webpage      │
└────────────────────────────────────────────────────┘

Component	Type	Role
Planner	LLM Chain	Analyzes query, creates research plan
Research Agent	ReAct Agent	Autonomously searches and gathers information
Report Generator	LLM Chain	Synthesizes findings into a Markdown report

Tools (used by the Research Agent)

Tool	Description
web_search	Search the web via Tavily API
get_search_urls	Get URLs from search results
fetch_webpage	Fetch full text content from a web page

Installation

Prerequisites

• Python 3.11+
• uv package manager

Setup

git clone <repository-url>
cd deep-research
uv sync
cp .env.example .env
# Edit .env with your API keys

Required API keys:

• OpenAI API key — for GPT-4
• Tavily API key — for web search
• Optional: Anthropic API key (for Claude as alternative)

Usage

Command Line

# Basic research
uv run python -m src.cli "What are the latest developments in quantum computing?"

# Use Anthropic
uv run python -m src.cli "Your query" --provider anthropic

# Verbose output
uv run python -m src.cli "Your query" --verbose

Python API

import asyncio
from src.main import DeepResearchSystem

async def main():
    system = DeepResearchSystem()
    result = await system.research("What are the latest developments in AI?")
    print(result["report"])

asyncio.run(main())

Project Structure

deep-research/
├── src/
│   ├── core/                  # Core components
│   │   ├── planner.py         # Research planner (LLM chain)
│   │   ├── researcher.py      # Research agent (ReAct with tools)
│   │   └── report_generator.py
│   ├── workflows/             # LangGraph workflow
│   │   ├── research_flow.py   # Plan → Research → Report pipeline
│   │   └── states.py
│   ├── tools/                 # Agent tools
│   │   ├── web_search.py      # Tavily web search
│   │   ├── content_fetch.py   # Web page content fetching
│   │   └── document.py        # Document chunking
│   ├── memory/                # Memory system
│   │   ├── working.py         # Current session memory
│   │   ├── short_term.py      # Recent sessions memory
│   │   └── manager.py
│   ├── llm/                   # LLM management
│   │   └── router.py          # Model router (OpenAI / Anthropic)
│   ├── config.py
│   ├── main.py
│   └── cli.py
├── tests/
├── examples/
├── pyproject.toml
└── .env.example

Development

uv run pytest           # Run tests
uv run black src/       # Format code
uv run ruff check src/  # Lint

License

MIT License