AI and the Web
The future of the web is now. Understand how AI and machine learning can be used by developers to build a more powerful web experience.
What is AI?
Understand the basics and definitions of the various emerging technologies, often referred to as AI.
Ethics and AI
As web practitioners, it's critical that we build new technology thoughtfully and responsibly.
Upgrade your site search
Discover techniques to get contextual results from generative AI, at-scale.
Practical tips
Understand LLM sizes
Take a look at a few real-world LLMs and the practical implications of different model sizes.
Compare model capabilities
Evaluate models and prompts with the LLM as a judge technique. Delegate validation to another LLM, instead of humans.
Improve performance and UX for client-side AI
Discover benefits such as low latency, reduced server-side costs, no API key requirements, increased user privacy, and offline access.
Better prompt engineering
Craft questions to get the best possible output from an LLM.
Download AI models with the Background Fetch API
Improve the user experience of downloading large AI models by implementing the Background Fetch API as a progressive enhancement.
Start building
Part 1: Client-side AI to combat online toxicity
Why toxicity detection matters and how you can use client-side AI to mitigate toxicity at its source.
Part 2: Build client-side AI toxicity detection
Learn how to build a client-side AI system to detect and mitigate toxicity at its source.
Product review suggestions
Online stores can see a 270% increase in conversions by displaying product reviews. Support better reviews with client-side AI.
Build a chatbot with WebLLM
WebLLM combines WebAssembly and WebGPU to perform on-device inference. Build a local and offline-capable chatbot in a classic to-do list application.
Build a chatbot with the Prompt API
Build a local and offline-capable chatbot in a classic to-do list application with an exploratory API from Google Chrome, the Prompt API.
Web AI Acceleration Fund
Chrome aims to foster AI innovation and adoption on the web.
By focusing on web-based AI solutions, we can democratize access to AI capabilities while leveraging the web's inherent strengths in distribution and deployment. We aim to accelerate innovation in areas that maximize these inherent advantages of the web platform.
Web AI Summit 2024
State of client side machine learning
See what's possible with client-side AI and where it's heading.
Discover Transformers.js
Learn about a JavaScript library that empowers developers to build never-before-seen web applications.
The Web Neural Network
See howtechnologies like WebAssembly and WebGPU have brought AI capabilities to the browser.
Gemini on the web
Learn how to access the Gemini API on the web, with the Google AI JavaScript SDK.
Generative AI samples
Discover example code and prompts for the Gemini API and Vertex AI API.
Responsible AI toolkit
Learn how to use AI responsibly, with best practices and resources.
AI on Chrome
Discover Google Chrome's documentation for AI, WebGPU, and WebAssembly.
Built-in AI APIs
Execute our most efficient model of the Gemini in Chrome, on-device.
Discover Console Insights
Ask Gemini to provide insights right in the DevTools Console, so you can better understand errors and warnings.
Use client-side AI
Client-side AI brings powerful models to users, while protecting data privacy and improving latency.
Supercharge Web AI model testing
Setting up a consistent testing environment with GPUs can be harder than expected. See the problems we faced and how we solved them, to improve your application performance.
Watch to learn
Watch these videos to learn AI fundamentals and discover TensorFlow JS.
Machine Learning for web practitioners
This introduction is aimed creatives, web developers, and JavaScript practitioners interested in using machine learning in their next project or idea.
AI, ML, and Deep Learning
Learn the foundations and definitions for some of the industry's most used buzzwords, and what they actually refer to.
What are pre-trained models?
Using pre-trained models can allow you to prototype an idea faster. In some cases, you can use these models in production.