Google’s mobile and web development platform is offering developers a new way to incorporate generative AI features into their applications. Available now in beta, Firebase Genkit is an open-source framework that blends diverse data sources, models, cloud services, agents and more with the coding style developers are used to.
“Genkit offers a rich AI-centric local developer tooling, making building and debugging your AI workload easier,” Google Product Manager Chris Gill and Developer Advocate Peter Friese write in a blog post. “Once you’re ready to go to production, you can use Genkit to deploy your solution to Firebase or Google Cloud and monitor your app to ensure it is production-ready.”
Though initially geared towards helping JavaScript/TypeScript developers make AI-powered apps available to Node.js backend developers, Google says it will soon add support for the Go programming language. Third-party open-source projects already supported by Genkit include vector databases like Chroma, Pinecone, Cloud Firestone and PostgreSQL; large language models from Ollama; and others with additional integrations planned over time.
The company claims Genkit can make developing AI features easier because it’s intuitive for developers, using familiar code-centric approaches. In addition, it comes out of the box with support for Gemini and Gemma. It’s also focused on local development, empowering developers to test their features end-to-end with full observability. Lastly, Genkit is open-sourced, flexible enough to handle plugins, capable of running seamlessly on Google Cloud infrastructure in leveraging Vertex AI, and is built for apps in production.
Some examples of what Genkit can do include generating content to produce writing based on a specified scenario or internal dataset, summarizing large texts for users in a simple and digestible form, producing high-quality responses grounded on proprietary data (RAG), providing text translation across languages, and extracting typed data from multimodal prompts.
“Learning new technologies takes time and effort, especially when it comes to AI with rapidly evolving concepts and approaches,” Gill and Friese state. “We believe AI frameworks should reduce complexity, not add to it. Genkit is designed to feel familiar and intuitive with a minimum learning curve to get started.”
Unveiled at this year’s Google I/O developer conference, Firebase Genkit is the company’s latest tool to facilitate AI-powered app development. Other products that have been infused with Gemini include Android Studio, Chrome DevTools, the Chrome desktop client, and its compliance platform app Checks. With GenKit, developers can move their AI-powered code from beyond the prototype and into production while understanding how the software is performing.
You can download Firebase Genkit from GitHub today.
The post Google introduces Firebase Genkit, a developer framework for building AI-powered apps appeared first on Venture Beat.
