LangGraph Research Agent

A local Python CLI research agent built with LangGraph, LangChain tools, OpenRouter, and Tavily.

What it does

• Runs an interactive chat loop in the terminal
• Uses a planner + tool-calling agent graph
• Searches the web and fetches specific pages
• Saves notes/reports to output/
• Persists conversation checkpoints in SQLite by thread ID

Requirements

• Windows PowerShell (or any shell)
• Python 3.10+
• API keys:
  • OPENROUTER_API_KEY
  • TAVILY_API_KEY

Quick start (PowerShell)

cd "C:\Users\Emil P\Desktop\Web Dev\langraph_agent"

# Create and activate a fresh virtual environment
py -m venv .venv
.\.venv\Scripts\Activate.ps1

# Install dependencies
pip install -r requirements.txt

# Configure environment variables
Copy-Item .env.example .env
# Then edit .env and add your real API keys

# Run the agent
python .\main.py

If python is not on PATH, use:

.\.venv\Scripts\python.exe .\main.py

CLI commands

• /help
• /verbose on
• /verbose off
• /threads
• /thread <id>
• /newthread
• /history [n]
• /save <filename>
• quit / exit / q

Files

• main.py: interactive CLI loop and user commands
• agent.py: LangGraph state machine, planner/chatbot/tool routing, checkpointing
• tools.py: web/search/file tools used by the agent
• output/agent_checkpoints.db: SQLite checkpoint storage (auto-created)

Notes

• save_to_file asks for approval by default in interactive sessions.
• Set AUTO_APPROVE_FILE_WRITES=1 in .env to skip save prompts.
• fetch_page now extracts readable text from HTML and returns source citations.