Deploy flows to any Node.js platform

Firebase Genkit has built-in integrations that help you deploy your flows to Cloud Functions for Firebase and Google Cloud Run, but you can also deploy your flows to any platform that can serve an Express.js app, whether it’s a cloud service or self-hosted.

This page, as an example, walks you through the process of deploying the default sample flow.

  1. Install the required tools:

    • Make sure you are using node version 20 or higher (run node --version to check).
  2. Create a directory for the Genkit sample project:

    export GENKIT_PROJECT_HOME=~/tmp/genkit-express-project
    mkdir -p $GENKIT_PROJECT_HOME
    cd $GENKIT_PROJECT_HOME

    If you're going to use an IDE, open it to this directory.

  3. Initialize a nodejs project:

    npm init -y
  4. Initialize a Genkit project:

    genkit init
    1. Select your model:

      Gemini (Google AI)

      The simplest way to get started is with Google AI Gemini API. Make sure it's available in your region.

      Generate an API key for the Gemini API using Google AI Studio. Then, set the GOOGLE_API_KEY environment variable to your key:

      export GOOGLE_API_KEY=<your API key>

      Gemini (Vertex AI)

      If the Google AI Gemini API is not available in your region, consider using the Vertex AI API which also offers Gemini and other models. You will need to have a billing-enabled Google Cloud project, enable AI Platform API, and set some additional environment variables:

      gcloud services enable aiplatform.googleapis.com
      export GCLOUD_PROJECT=<your project ID>
      export GCLOUD_LOCATION=us-central1

      See https://cloud.google.com/vertex-ai/generative-ai/pricing for Vertex AI pricing.

    2. Choose default answers to the rest of the questions, which will initialize your project folder with some sample code.

  5. Build and run the sample code:

    npm run build
    genkit flow:run menuSuggestionFlow "\"banana\"" -s
  6. Optional: Start the developer UI:

    genkit start

    Then, navigate to http://localhost:4000/flows and run the flow using the developer UI.

    When you're done, press Ctrl+C in the console to quit the UI.

  7. Try out the Express endpoint:

    npm run start

    Then, in another window:

    curl -X POST "http://127.0.0.1:3400/menuSuggestionFlow?stream=true" -H "Content-Type: application/json"  -d '{"data": "banana"}'
  8. If everything's working as expected, you can deploy the flow to the provider of your choice. Details will depend on the provider, but generally, you need to configure the following settings:

    Setting Value
    Runtime Node.js 20 or newer
    Build command npm run build
    Start command npm run start
    Environment variables GOOGLE_API_KEY=<your-api-key> (or whichever secrets are required)