LangGraph - Agent from template
pip install "langgraph-cli[inmem]"
If working within a monorepo, navigate to the packages directory.
Create the Poetry package.
poetry new --src $"{APP_NAME}"
Create a react agent from the template in a temporary location, e.g., tmp.
langgraph new tmp --template react-agent-python
Copy all of the files into the APP_NAME directory, except pyproject.toml.
Rename
react_agentRename all references to react_agent, the default package name, to your desired package name.
For example:
- Rename the
src/react_agentdirectory tosrc/app_name - Rename the
react_agentin langgraph.json to the app name.
Add the package dependencies.
poetry add langgraph langchain-openai langchain-anthropic langchain langchain-fireworks python-dotenv langchain-community tavily-python
poetry add --group dev mypy ruff
Install the LangGraph CLI locally.
poetry add --group dev "langgraph-cli[inmem]"
Create the .env file.
cp .env.example .env
Add any keys to the .env file, e.g., OpenAI.
Run LangGraph Studio.
poetry run langgraph dev
Resources
- https://studio.langchain.com/
- Quickstart: Launch Local LangGraph Server, i.e., LangGraph Studio
LangGraph - Agent from template
Interactive graph
On this page
Resources