This is a concise guide on setting up and running the LangGraph development server (`langgraph dev`) using `uv` as the package manager and runner.

langgraph_uv_guide.md

Running LangGraph Dev with UV

This is a concise guide on setting up and running the LangGraph development server (langgraph dev) using uv as the package manager and runner.

Prerequisites

1. Python: Ensure you have Python 3.11+ installed.
2. UV: Install uv: curl -LsSf https://astral.sh/uv/install.sh | sh (or other methods from astral.sh/uv).
3. .env file: Have your necessary API keys (e.g., OPENAI_API_KEY) in a .env file in your project root.
4. LangGraph Agent: Have your LangGraph agent code (e.g., in agent.py) defining a compiled graph (e.g., graph).

Setup

1. langgraph.json: Create a langgraph.json file in your project root. This file tells the LangGraph CLI where to find your graph(s).

   {
     "dependencies": ["."], // Optional: If your agent has local dependencies
     "graphs": {
       "agent": "./agent.py:graph" // Path to your file and graph variable
     },
     "env": ".env" // Points to your environment variables file
   }

   • Replace "./agent.py:graph" with the actual path to your agent file and the variable name of your compiled graph.

2. Install Dependencies: Use uv to install langgraph-cli with the necessary extras for the dev server, along with any other project dependencies (e.g., langchain-openai, python-dotenv).

   # Activate your virtual environment if you have one (optional but recommended)
   # uv venv
   
   # Install langgraph-cli and other dependencies
   uv pip install --upgrade "langgraph-cli[inmem]" langchain-openai python-dotenv # Add other dependencies here

   • The [inmem] extra is required for langgraph dev.

Running the Server

1. Start the Dev Server: Use uv run to execute the langgraph dev command within the environment managed by uv. Navigate to your project root in the terminal and run:

   uv run langgraph dev

2. Output: You should see output indicating the server is running, including URLs for the API, docs, and the LangGraph Studio:

   Ready!
   * API: http://localhost:2024
   * Docs: http://localhost:2024/docs
   * LangGraph Studio Web UI: https://smith.langchain.com/studio/?baseUrl=http://127.0.0.1:2024
   

Accessing LangGraph Studio

• Open the LangGraph Studio Web UI link in your browser (Note: Safari might have issues with local servers) to visualize, interact with, and debug your agent. The baseUrl parameter connects the web UI to your locally running server.