{"slug": "15-updates-from-google-i-o-2026-powering-the-agentic-web-with-new-capabilities", "title": "15 updates from Google I/O 2026: Powering the agentic web with new capabilities, tools, and features in Chrome", "summary": "At Google I/O 2026, Google unveiled a vision for the \"agentic web,\" introducing 15 updates to Chrome focused on empowering AI agents, enhancing web UI and performance, and integrating Gemini as a proactive browser assistant. Key announcements included WebMCP, a proposed open standard for exposing website functions to agents, and Modern Web Guidance, a tool to help coding agents build modern, accessible web experiences. These updates aim to make web browsing smarter and more efficient by bringing built-in AI models and automation tools directly to the browser.", "body_md": "Published: May 19, 2026\n\nAgents are transforming development everywhere, and nowhere is that transformation happening faster than on the web. It's redefining what we build, how we build, and who builds. As we enter the era of the agentic web, we see a shift that bridges the gap between complex developer workflows, underlying platform capabilities, and everyday user experiences.\n\nAt Google I/O 2026, we unveiled a vision for this era. These ideas tie\ntogether three core areas of the web ecosystem: **empowering AI agents** to\nbuild and interact with websites through new capabilities,\n**pushing the boundaries of web UI and performance**,\nand transforming the browser into a\n**powerful, proactive assistant** **for everyday users with Gemini in Chrome**.\nBy integrating efficient, built-in AI models directly to the browser and\nbringing powerful automation tools like auto browse to Chrome, we're making using the\nweb smarter, faster, and more accessible for everyone.\n\nHere are the 15 biggest updates we shared at Google I/O 2026 to help you build and thrive in the agentic era of web browsing.\n\n## Empowering AI agents for the Web\n\nAgents are changing how we interact with software, and we believe the web must be equipped to guide them. We are introducing powerful new capabilities and tools, like WebMCP and Modern Web Guidance, that allow you to build modern web experiences with greater clarity and velocity. We're also giving you the AI-assisted tools you need to build, debug, and optimize code faster and more accurately than ever before.\n\n### 1. WebMCP: Transform your websites into agentic toolkits\n\nWe're giving you a first look at\n[WebMCP](/docs/ai/webmcp),\na proposed open web standard that lets you expose structured tools like JavaScript\nfunctions and HTML forms to browser-based agents. By defining these tools,\nyou can instruct agents exactly how and where to interact with your site.\nThe result? An agent can now call machine-friendly functions to complete\ncomplex tasks in seconds with greater reliability, precision, and\npersonalization. Imagine a user is planning a multi-city vacation. Instead\nof watching an agent click through travel forms, they can authorize it to\nquery backend APIs directly to instantly build a personalized,\nweather-optimized itinerary for their approval.\n\nThe experimental WebMCP origin trial starts in Chrome 149. Gemini in Chrome will soon support WebMCP APIs. We're already seeing global consumer brands experimenting with WebMCP to create more delightful and engaging experiences for their users.\n\n### 2. Modern Web Guidance: A blueprint to guide coding agents to build for the modern web\n\n[Modern Web Guidance](/docs/modern-web-guidance), now available in\nearly preview, is a set of evergreen and expert-vetted skills that guide\nyour coding agents across many common use cases to build modern web\nexperiences that are the most accessible, performant, and secure. It\nintegrates directly with [Baseline](http://web.dev/baseline), letting you\nfocus on what you want to build while your tools automatically figure out\nthe right features and fallbacks to use within your chosen Baseline target.\nInstall with a single click in Google Antigravity, through npx or as an\nextension in a coding agent. Modern Web Guidance features support for over\n100 use cases for dozens of the latest features with continuous updates\nadded regularly.\n\n### 3. Automate debugging with Chrome DevTools for agents\n\nScale your workflow with [Chrome DevTools for\nagents](/blog/devtools-for-agents-v1), which provides\nvisibility to verify, debug, and optimize code in real time. By providing\nagents with direct access to DevTools' capabilities, such as console logs,\nnetwork traffic, and accessibility trees, they can verify and automate fixes\nwithout manual oversight.\n[Chrome DevTools for agents](/docs/devtools/agents) is available\ntoday for Antigravity and more than 20 other coding agents.\n\n### 4. Gain deep insights with AI-assisted debugging in Chrome DevTools\n\n[AI assistance in Chrome DevTools](/docs/devtools/ai-assistance/get-started)\nnow has access to Lighthouse data, and can automatically search for context\nto answer more open-ended questions than were previously possible.\nAdditionally, widgets give you full visibility into Gemini's reasoning to\nhelp you with debugging.\n\n### 5. Skip servers, budgets, and red tape: Unlock AI features with built-in AI\n\nRunning entirely in the browser, built-in AI lets you deploy personalized, proactive features that would be cost-prohibitive on the server. Skip the token bills and other obstacles to focus entirely on unique user value. Best of all, the browser manages and shares optimized models across sites, enabling more users to enjoy AI experiences on the web.\n\nTo help you build these frictionless AI enhancements, we're expanding the web AI toolkit:\n\nChrome 148 uses Gemini Nano with multimodal inputs and structured output for rich experiences, reliable JSON for seamless integrations, and access to expanded language support.[Prompt API](/docs/ai/prompt-api)is stable:**Gemma 197M:** This ultra-efficient expert model can transparently power task-specific APIs like*summarizer*, automatically scaling your features to a broader spectrum of devices.\n\n[Explore the full built-in AI suite](/docs/ai/built-in-apis), including our\nexisting Translator and Language Detector APIs, and\n[join the Early Preview Program](https://goo.gle/chrome-ai-dev-preview-join)\nto test upcoming APIs.\n\n## Pushing the boundaries of web UI and performance\n\nWe're developing next-generation platform features that continue to blur the line between web and native apps. New declarative APIs, such as HTML-in-Canvas and Declarative Partial Updates, handle complex rendering and performance tasks for you, making it easier than ever to build beautiful, modern, high-fidelity, performant and interactive experiences on the web.\n\n### 6. HTML-in-Canvas and element-scoped view transitions: break the boundaries with next-gen UI\n\nThe new [HTML-in-Canvas API](/blog/html-in-canvas-origin-trial) and\n[element-scoped view transitions](/docs/css-ui/view-transitions/element-scoped-view-transitions)\nenable previously impossible UIs that bring high-fidelity, app-like\ninteractivity to the web. With the HTML-in-Canvas API, integrate real DOM\nelements directly into a canvas with WebGL and WebGPU to build an\n[immersive 3D experience](https://github.com/GoogleChromeLabs/css-web-ui-demos/blob/main/html-in-canvas/awesome-html-in-canvas.md)\nthat is searchable, accessible, natively translatable, and interacts\nseamlessly with your built-in browser features. Combine this with view\ntransitions—like element-scoped, available now in Chrome 147, and two-phase\ntransitions, currently in testing—to create layered UI motion and animate\nintermediate states without blocking page interactivity. By turning complex\ninteractions into declarative APIs, we deliver high-fidelity performance by\ndefault. The\n[HTML-in-Canvas API origin trial](/origintrials#/register_trial/3478467762190286849)\nis available now.\n\n### 7. Performance and UI wins: Core Web Vitals for SPAs and more\n\nChrome is enabling new ways to improve performance for modern app-like web\nexperiences. New updates include the\n[Soft Navigations API](/docs/web-platform/soft-navigations),\navailable in an upcoming Chrome release, to bring Core Web Vitals measurement to Single Page\nApplications. We're also introducing new\n[Declarative Partial Updates](/blog/declarative-partial-updates)\nprimitives bringing native out-of-order HTML updates to the platform as well\nas new streaming APIs to make it easier to insert HTML into the page without\nheavy DOM manipulation. These APIs are available for testing now.\n\n### 8. Modernize authentication with Immediate UI mode\n\nAs part of our identity updates, Immediate UI mode unifies passwords and\npasskeys into a single, browser-managed sign-in flow. When a user clicks\n\"Sign In\" on your site, Chrome automatically surfaces the available\ncredentials—allowing for seamless authentication using saved passwords or\npasskeys. Get started with the\n[Immediate UI mode implementation guide](/docs/identity/immediate-ui-mode).\n\n### 9. Plan your Baseline target with real-world traffic data\n\nNo more shuffling data around with exported TSV files! Use the new [Baseline Checker](https://baseline-checker.chrome.dev/)\ntool to connect directly to the updated Google Analytics API and see exactly\nwhat percentage of your actual users support modern features. Pick a\nBaseline target and confidently ship the latest features to your users,\nwhile knowing when to use fallbacks.\n\n## Supercharging the browsing experience with Gemini in Chrome\n\nWith Gemini in Chrome on desktop, iOS and now Android, we're giving users powerful new ways to browse, create, and get things done. From automating complex, multi-step tasks with auto browse to intuitive multimodal interactions using your cursor or voice, Gemini in Chrome puts powerful productivity directly at the user's fingertips.\n\n### 10. Gemini in Chrome for Android: A browsing assistant on your phone.\n\nComing in June, we designed Gemini in Chrome on Android to be your personal\nbrowsing assistant, helping you better understand\ncontent on the web. It lets you summarize long articles, ask specific\nquestions, and get detailed explanations without having to switch apps.\nBeyond answering questions, it acts as a versatile productivity tool that\nconnects with Google apps like Calendar, Keep, and Gmail to help you\ncomplete tasks quickly. And with\n[Personal Intelligence](https://blog.google/products-and-platforms/products/search/personal-intelligence-expansion/),\nif you choose to connect apps like Gmail and Google Photos, this secure,\ncontext-aware browsing assistant can even provide tailored responses based\non your unique interests, hobbies and more.\n\n### 11. Use auto browse to take care of tedious tasks\n\nAlready available on desktops,\n[auto browse for Android](http://blog.google/products-and-platforms/products/chrome/bringing-chrome-ai-to-android/)\nlets you get the most out of Gemini in Chrome by automating your digital chores\nso you can focus on more important\ntasks. With auto browse, you can easily complete tasks from appointment\nbooking to party planning, finding in stock items and more, all from your\nAndroid phone. For example, if you are about to head out to a comedy show,\nbut forgot to book parking, auto browse has you covered. Just ask Gemini in\nChrome, and it will gather event details from your ticket to find you a spot.\n\nOn desktop, we will be integrating auto browse with Gemini Spark in the coming months, so that your 24/7 personal AI agent can take actions in the browser on your behalf.\n\n### 12. Transform images on the go with Nano Banana\n\nWith Nano Banana, you can instantly create or customize images while browsing the web on your Android device. Just ask Gemini in Chrome to \"turn this page into an informative infographic\" while studying, or \"alter the image to include modern living room essentials\" when browsing for apartments.\n\n### 13. Skills in Chrome: Turn your best AI prompts into one-click tools\n\nSkills in Chrome let you save and reuse your most helpful AI prompts in Gemini in Chrome on desktop. Save a multi-tab workflow once, like generating side-by-side spec comparisons while shopping or scanning long documents for key information, and run it again instantly with a single click any time.\n\n### 14. Select from your screen to prompt Gemini in Chrome\n\nYou can now use your mouse pointer to ask Gemini in Chrome about the specific parts of the webpage you are looking at, saving you from having to describe exactly what you mean. For example, you can select two products on a page and instantly compare their key features. Or if you want to edit an image with Nano Banana you can select exactly the part of the image you want to change\n\n### 15. Use your voice across the web\n\nSoon it will be possible to use your voice to type into websites across Chrome on desktop. With voice, it will be easier and more natural to do things like drafting comments, filling in long fields on forms, or writing emails. This will use Gemini models to either clean up your transcription - removing the ums and ahs and fitting it to the context while keeping true to your voice - or fill in the field as you ask it to.\n\n## What's next\n\nThe transition to the agentic web is unfolding right in front of us. By bridging the gap between powerful underlying AI capabilities and everyday web development, we are removing the friction that has historically slowed down and constrained innovation.\n\nWe're moving away from a web that requires you to do all the heavy lifting, and\ntowards a web that proactively works *for* you. Whether you are looking to\nseamlessly integrate with browser-based agents, push the absolute visual limits\nof what a webpage can do, or simply streamline your own debugging workflow,\nwe're bringing you the tools to build the future of the web.\n\nExplore the full guides and technical deep-dives at\n[developer.chrome.com](/) and\n[web.dev](https://web.dev/) We want to see what you build with these new\nfeatures. So share your projects and be sure to connect with us\n[X](https://twitter.com/ChromiumDev),\n[LinkedIn](https://www.linkedin.com/showcase/chrome-for-developers/), and\n[YouTube](https://www.youtube.com/user/ChromeDevelopers).\n\nSee you at the next Google I/O!", "url": "https://wpnews.pro/news/15-updates-from-google-i-o-2026-powering-the-agentic-web-with-new-capabilities", "canonical_source": "https://developer.chrome.com/blog/chrome-at-io26?hl=en", "published_at": "2026-05-19 07:00:00+00:00", "updated_at": "2026-05-24 05:06:55.652955+00:00", "lang": "en", "topics": ["artificial-intelligence", "developer-tools", "products"], "entities": ["Google", "Chrome", "Gemini", "Google I/O 2026", "WebMCP", "Modern Web Guidance"], "alternates": {"html": "https://wpnews.pro/news/15-updates-from-google-i-o-2026-powering-the-agentic-web-with-new-capabilities", "markdown": "https://wpnews.pro/news/15-updates-from-google-i-o-2026-powering-the-agentic-web-with-new-capabilities.md", "text": "https://wpnews.pro/news/15-updates-from-google-i-o-2026-powering-the-agentic-web-with-new-capabilities.txt", "jsonld": "https://wpnews.pro/news/15-updates-from-google-i-o-2026-powering-the-agentic-web-with-new-capabilities.jsonld"}}