This is a demonstration of LangChain with Google's Gemini Pro model. Not intended for production use.

After cloning the repo, grab a Gemini API key from here and export it as an environment variable

export GOOGLE_API_KEY=<YOUR-API-KEY>

Create a virtual environment and install the requirements

python -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Run the application server

python -m app.main

Open the playground in your browser at http://localhost:5000/

If you don't have a Google Cloud account, you can get one for free with $300 in credits. You will also need a local Docker and gcloud CLI.

Set your default Google Cloud project

export GOOGLE_PROJECT_ID=<YOUR-PROJECT-ID>
gcloud config set project $GOOGLE_PROJECT_ID

Build and push the image to Google Cloud

export GEMCHAIN_IMAGE="gcr.io/<YOUR-PROJECT-ID>/gemchain:latest"
gcloud docker -- push $GEMCHAIN_IMAGE

Deploy to Google Cloud Run and allow unauthenticated access

gcloud run deploy gemchain \
    --allow-unauthenticated \
    --image $GEMCHAIN_IMAGE \
    --platform managed \
    --region us-central1 \
    --set-env-vars GOOGLE_API_KEY=$GOOGLE_API_KEY

MIT