Google’s Firebase gets AI extensions, opens up its marketplace

Firebase, Google’s backend-as-a-service platform for application developers, is getting its fair share of new features at the company’s I/O developer conference today, but the marquee update is the addition of AI extensions, powered by Google’s PaLM API. That’s in addition to support for the Python language in Cloud Functions for Firebase (2nd gen), updates to the Cloud Firestore database service and more. In addition, Google is now opening up the Firebase extension marketplace to more developers.

Jeanine Banks, the lead for Google’s Developer Product group, told me that the company is focused on reducing complexity for developers. “With so many different technologies — especially now that you want to build for multiple devices, you want to build across multiple platforms — certainly we’re not short on tools to help developers do that. So it’s critical that we make it as easy as possible to build great things,” she said.  “Of course, with AI and generative AI, that just ups the stakes again. It’s more for developers to think about and in particular, not just those developers that are deep into AI and machine learning, but the everyday engineer. How can they take advantage of these things to build great apps and experiences?”

Google first introduced Firebase extensions, which are pre-packaged solutions for app developers, back in 2019. The service is generally available now and given Google’s focus on all things AI right now, it’s no surprise that the Firebase team — and the developers who use the platform — are jumping on this bandwagon, too. Among the new extensions here are image generators, tools for building chatbots on top of the PaLM API and a text summarizer, among others.

Until last year, Google was working directly with the extension providers that wanted to offer their APIs in the marketplace. Now, it is opening up this project to more providers so that they can onboard themselves, Banks told me.

“Our main focus with Firebase extensions is to accelerate developers’ creativity and being able to build their applications,” she explained. “We think by enabling them to that, they will succeed, their applications will succeed — and as their applications succeed, they will likely scale up. They’ll use more cloud, more Play Services, they’ll use more payment services.” That’s also how Google plans to (indirectly) monetize this service — not through revenue of profit sharing arrangement but by simply getting more developers into the larger ecosystem of Google services.

The addition of Python as a supported language in Cloud Functions, too, is somewhat related to this AI play. Python, after all, is the default language for most AI projects and as Banks noted, this will now allow developers in the AI/ML community to build their own extensions more easily and make them available in the marketplace as well.

Specifically, Cloud Functions 2nd gen, which is also now generally available, is where Firebase is adding this Python support (in public preview). In addition, Cloud Functions 2nd gen, now offers up to 32 gigabytes of memory and can handle up to 1000 requests in parallel.

Some of the smaller updates to Firebase this year include support for OR queries in Firestore and a number of security updates, as well as support for frameworks like SvelteKit, Asto and Nuxt in Firebase Hosting, and support for Fultter SDK updates and Firebase Hosting support for Flutter web using WebAssemby.