Implementation of a customer support agent that collects user information, validates it, and only then assists the user. The state is persisted to allow continuation and provide a more advanced user experience.

• Human In the Loop: Continuously gather input from the user until the agent has all required information.
• Persistent Storage: Checkpoint LangGraph state after each node execution to an SQLite database.

To run this project, you will need to add the following environment variables to your .env file

PYTHONPATH=/{YOUR_PATH_TO_PROJECT}/langgraph-customer-support

MISTRALAI_API_KEY

LANGCHAIN_API_KEY

LANGCHAIN_TRACING_V2=true

LANGCHAIN_PROJECT

Clone the project

  git clone https://github.com/emarco177/langgraph-customer-support.git

Go to the project directory

  cd langgraph-customer-support

Install dependencies

  poetry install

Start the flask server

  poetry run app.py

To run tests, run the following command

  poetry run pytest . -s -v