Add AI features to your app using Gemini CLI and Firebase AI Logic

We’re excited to announce our new Firebase extension for Gemini CLI, which provides a powerful, integrated way to work with your Firebase projects and codebase directly from your terminal. In addition to automating Firebase setup, it also streamlines feature development, allowing Gemini CLI to implement AI features in your app, such as analysis of multimodal input and nano banana image generation directly in your application code for you.

Start adding AI features to your apps

The new Firebase extension for Gemini CLI comes bundled with the Firebase MCP server. It includes a new /firebase:init command which helps the Gemini CLI set up Firebase AI Logic and add the AI feature into your actual codebase. We’re launching with support for web, Flutter, and Android apps to ensure a wide range of developers can take advantage of this new agentic workflow.

Adding an AI feature to your mobile and web applications is now as simple as three steps:

  1. Install the Firebase extension for Gemini CLI.
  2. Ask the CLI to set up your project to use AI Logic and the Gemini Developer API.
  3. Describe to the Gemini CLI which AI feature you want to implement.

Some best practices using /firebase:init:

  • Since Gemini is non-deterministic, account for some variation in the experience across sessions. You can remind it to check its instructions if it makes mistakes.
  • If you encounter any errors, share them with Gemini CLI so it can help fix them.
  • Frequently commit snapshots of your code, especially before making major changes to your app.
  • Test your code before deploying to production.

Walking through the experience

Install the Firebase extension for Gemini CLI

Make sure that you have the latest version of Gemini CLI installed. Then, install the new Firebase extension, which provides the Gemini CLI with access to the Firebase MCP tools and prompts.

gemini extensions install https://github.com/gemini-cli-extensions/firebase

Ask the CLI to set up your project to use AI Logic and the Gemini Developer API

Next you trigger the /firebase:init prompt and select the Firebase AI Logic option. Gemini CLI sets up a Firebase project for you, registers apps in your Firebase project based on your platform, enables the Gemini Developer API, and adds the necessary SDK initialization code to your application.

image1
image1
image2
image2

Describe to Gemini CLI which AI feature you want to implement

Lastly, you describe – in natural language – the AI feature you want to implement, sharing any relevant context about where the feature should be added into your app and how it should behave. The /firebase:init prompt provides the Gemini CLI with applicable code snippets for AI Logic so that Gemini can combine them with its knowledge of your codebase to implement your requested feature.

image3
image3
image4
image4

What’s next?

Our commitment is to continually enhance your development journey. We’re excited to see what innovative AI features you’ll build with the new Firebase extension for Gemini CLI.

Your feedback is crucial as we continue to evolve. When you try out the new extension, please let us know what you think in our GitHub repository.

Also check out the Firebase AI Logic docs for comprehensive information and guidance on all the capabilities that the product supports.