The application is composed of two parts - the api and the ui.
Switch to the api directory. All commands below are to be run there.
uv venv source .venv/bin/activate uv pip install -r requirements.txtThis should get all the required packages set up. Now, let's set the environment.
Head to Google AI Studio to get a Gemini API Key.
And grab a Weights and Biases key from https://wandb.ai/authorize.
Now, set the following in the .env file:
GEMINI_API_KEY="YOUR_KEY_HERE" GOOGLE_GENAI_USE_VERTEXAI=false You should now be able to run the API endpoint locally by:
uvicorn main:app --reloadThis should bring up the API at http://localhost:8000.
Now, switch to the ui directory from the project root in a new terminal tab/window.
pnpm iThis should be enough to install all the requirements. Now we can run it with
pnpm devThat should do the trick for launching the UI on your local browser at http://localhost:3000.