Python AI agent with LangGraph project structure

Python AI agent with LangGraph project structure

Explorer

  • .venv…, not committed
  • .langgraph_api…, not committed
  • langgraph.json
  • pyproject.toml
  • uv.lock
  • .env.example
  • .env, not committed
  • .gitignore
  • .codespellignore
  • Makefile
  • README.md
  • LICENSE
  • static
  • studio_ui.png
  • src
  • tests
  • conftest.py
  • cassettes…new
  • evalsnew
  • evaluators.pynew
  • run.pynew
  • sync_datasets.pynew
  • README.mdnew
  • results…new, not committed
  • .vscodenew
  • settings.jsonnew
  • launch.jsonnew
  • scriptsnew
  • chat.pynew
  • .pre-commit-config.yamlnew
  • AGENTS.mdnew
  • .github

  1. # Install the LangGraph CLI with the in-memory dev server (needs Python 3.11+)

    uv tool install "langgraph-cli[inmem]"

  2. # Create the app from the official template

    langgraph new support-agent --template new-langgraph-project-python

  3. # Install dependencies and copy the env file

    cd support-agent && uv sync && cp .env.example .env

  4. # Run the dev server and open LangGraph Studio

    uv run langgraph dev

  5. # Growing: add a model provider, tools and test helpers

    uv add langchain langchain-anthropic httpx && uv add --dev pytest-recording agentevals openevals

  6. # Production: build the server image from langgraph.json

    uv run langgraph build -t support-agent:latest

  7. # Production: run the image locally with its database and queue

    uv run langgraph up