Alphabet (GOOG) Google I/O '25 Developer Keynote summary

Key product announcements and demos

• Introduced Gemini 2.5 Pro and Flash, highlighting their integration into coding agents and rapid prototyping tools.

• Showcased Sketch, an experimental tool for generating app designs from prompts, with direct export to Figma and code.

• Demonstrated AI Studio's new features, including native audio support, URL Context for web grounding, and improved function calling.

• Launched updates to Google Gen AI SDK with MCP support and showcased agentic apps using Google Maps.

Developer tools and productivity enhancements

• AI Studio now features a native code editor optimized for Gemini 2.5 Pro, enabling rapid, iterative app development.

• Android Studio integrates Gemini for natural language-driven testing, dependency updates, and code refactoring.

• Gemini Code Assist is now available for enterprise, offering privacy and security for business environments.

• Firebase Studio enables full-stack app creation from prompts, Figma imports, and automated backend provisioning.

AI and multimodal advancements

• Gemini models now support 24 languages, proactive audio, and session context management.

• Multimodal capabilities allow apps to process text, images, audio, and video, with seamless integration into Android and web.

• Gemini Nano powers on-device AI for summarization, rewriting, and image description, ensuring privacy.

• New multimodal APIs in Chrome and Firebase enable hybrid on-device/cloud AI for web and mobile apps.