Google, at its I/O 2025 conference significantly enhanced its Firebase developer platform with new AI-powered tools, notably launching Firebase AI Logic and the experimental Stitch UI generator. These introductions aim to simplify AI integration and accelerate app creation, reflecting Google’s strategy to embed generative AI across the development lifecycle and cater to the “vibe coding” trend. This move is poised to make sophisticated AI more accessible to a broader range of developers.
Firebase AI Logic, an evolution of Vertex AI in Firebase, offers a toolkit for integrating Google’s Gemini models into apps client-side or via the server-side Genkit framework, as detailed in a Firebase Blog post. The platform now provides client-side access to the Gemini Developer API, currently in Preview with a no-cost tier, which the Firebase Blog describes as “the easiest way to get started with generative AI.” This runs alongside the generally available Vertex AI Gemini API, which remains the option for enterprise-grade needs.
Further client-side enhancements include experimental hybrid on-device (Gemini Nano on Chrome) and cloud inference (also in Preview), a Preview SDK for Unity, Gemini-powered image generation, and improved AI monitoring dashboards. For server-side work, Genkit for Node.js now supports dynamic model lookup for the latest Gemini models, a feature DevClass highlighted as beneficial for reducing friction in adopting AI advancements. Google is also broadening Genkit’s language support to include Go (Beta) and Python (Alpha) Genkit Go and Python support.
Stitching UI Design With AI
Complementing Firebase AI Logic, Google Labs introduced Stitch, an experimental tool that uses Gemini 2.5 Pro’s multimodal capabilities to generate UI designs and frontend code from text or image prompts. Stitch aims to optimize workflows between design and development by creating visual interfaces from natural language or uploaded wireframes. Users can choose between Gemini 2.5 Pro and Gemini 2.5 Flash models. Stitch supports rapid iteration, allows pasting designs directly into Figma, and exports functional frontend code.
Kathy Korevec, a Google product manager, told TechCrunch that Stitch is where developers “can come and get your initial iteration done, and then you can keep going from there.” She further elaborated t on their vision, stating their aim is to make this next level of design thinking and software building “super, super easy and approachable for people… to do that next level of design thinking or that next level of software building for them.” However, Stitch is not intended to be a full-fledged design platform like Figma or Adobe XD. While Stitch is quick for first drafts, complex designs benefit from the Figma export for detailed refinement. Stitch’s use of Gemini 2.5 Pro makes it a noteworthy tool for accelerating initial web app prototyping.
Embracing ‘Vibe Coding’ And Broader Tooling Enhancements
These announcements align with the growing “vibe coding” trend, an intuitive, AI-assisted method of software creation. Jeanine Banks, VP and General Manager of Developer X at Google, remarked to SiliconANGLE that “Maybe a year ago, we weren’t imagining that even professional developers, people who code, would want to have sort of the vibe experience of coding, and that’s really exploded,” and that this type of coding, she added, has “grown and… [is] allowing people who don’t know how to code to be able to get started, be able to prototype.”
She also stated that with Firebase AI Logic, Google is “combining [AI capabilities] and enhancing them so developers can have a one-stop shop to integrate all kinds of AI models.”
This strategy extends to Firebase Studio, which launched on April 9th. At I/O, Firebase Studio gained Figma import support via a partnership with Builder.io and improved backend integration recommendations. Firebase Studio itself, a Code OSS-based environment, merged Project IDX and features an “App Prototyping agent.” While these tools aim to speed up development, it will output occasional generation errors and make vibe-coders more dependent on Google.
Google also expanded access to Jules, its AI agent for bug fixing and code upgrades, now in public beta and utilizing Gemini 2.5 Pro. These offerings position Google within a competitive landscape that includes tools like GitHub Copilot, Cursor, and OpenAI’s Codex.
