Google has launched Stitch, an AI-powered tool for designing application front-ends, at its I/O 2025 conference. This introduction marks Google’s significant step into the fiercely competitive arena of AI-assisted design and software development. Stitch leverages the company’s Gemini 2.5 AI models to generate user interface (UI) elements along with corresponding HTML and CSS code directly from text or image prompts.
The new tool, accessible at stitch.withgoogle.com, is designed to substantially streamline the workflow between initial design concepts and functional code. Google’s official announcement explains that Stitch, an experiment from Google Labs, empowers users to rapidly transform simple inputs into complex UI designs and frontend code. Developers and designers can then export these AI-generated creations to Figma for further refinement or directly utilize the exposed code within their preferred Integrated Development Environments (IDEs).
Stitch: Bridging Design and Code with AI
Users of Stitch can select between Google’s Gemini 2.5 Pro and Gemini 2.5 Flash AI models to power the interface ideation and code generation. Google product manager Kathy Korevec described Stitch as a place “where you can come and get your initial iteration done, and then you can keep going from there,” and emphasized the company’s aim to make the next level of design and software building “super, super easy and approachable” for users.
However, Google clarifies that Stitch is not intended as a full-fledged replacement for comprehensive design platforms like Figma or Adobe XD, but rather as a tool to accelerate the initial stages of UI development.
An upcoming feature will further enhance this by allowing users to make UI changes through annotated screenshots.. Google also highlights that Stitch facilitates conversational iteration on designs and allows for easy theme adjustments.
Navigating the Crowded AI Coding Landscape
Stitch enters an increasingly dynamic “vibe coding” market, where AI tools are reshaping software creation. OpenAI recently made waves with the launch of Codex, an AI coding agent integrated into ChatGPT, designed to automate a variety of software development tasks. Microsoft, a key OpenAI partner, also continues to enhance its widely used GitHub Copilot.
The competitive field is further populated by innovative startups. These include Anysphere, the company behind Cursor; Cognition, which developed the Devin AI coding agent; and Windsurf.
Windsurf, which OpenAI is reportedly in the process of acquiring for approximately $3 billion, recently launched its own SWE-1 AI model family, with Windsurf’s CEO Varun Mohan emphasizing that “writing code is only a fraction of what engineers do.”
Adding to this competitive mix, Zencoder recently introduced Zen Agents, a platform offering customizable AI coding agents and an open-source marketplace.
Google’s Expanding AI Developer Toolkit
Beyond the introduction of Stitch, Google has broadened access to Jules, its AI coding assistant. Jules is powered by Gemini 2.5 Pro and offers robust capabilities. Google explains that Jules can address bug backlogs, manage multiple tasks simultaneously, and initiate new feature development.
It integrates directly with GitHub, cloning repositories to a Cloud VM and creating pull requests for project merging. This development aligns with a significant industry trend where AI tools are increasingly responsible for substantial portions of new code, a point noted by some tech leaders.
The Gemini 2.5 Pro model itself now features an experimental Deep Think mode for tackling complex mathematical and coding problems. Complementing these tools, Google AI Studio’s native code editor has been optimized for the Gemini SDK, even enabling the self-correction of code errors.
In a move to empower more developers, Google is also extending Gemini Nano access to third-party Android developers through new ML Kit GenAI APIs, which will facilitate on-device AI features like summarization and image description for enhanced privacy and offline functionality.
This comprehensive strategy reflects the high stakes in the AI race, as Google CEO Sundar Pichai remarked at I/O 2025, “The opportunity with AI is truly as big as it gets.”
While the advancements promise accelerated development, the broader developer community maintains a degree of caution. Discussing OpenAI’s Codex, open-source contributor DevChampion stated his view that “the community must remain vigilant about its impact on fundamental skill acquisition for those entering the field.”
OpenAI itself underscores the importance of human oversight for its Codex tool, stating, “It still remains essential for users to manually review and validate all agent-generated code before integration and execution” – a critical reminder applicable to all AI-generated code.
