Use Web APIs to Build Browser AI Apps

AI is thriving on the server side, but the web is catching up! Let’s explore how to build browser AI apps, like an offline speech-to-speech translator, leveraging current and future Web APIs.

02:12 - The AI Revolution on the Web 03:44 - Building an Offline Translation App 04:55 - Web Speech API: Speech Recognition & Synthesis 07:06 - Large Language Models for Translation 08:28 - Demo: Speech Recognition & Synthesis in the Browser 11:12 - The Challenge of Translation in the Browser 13:20 - AI Models in the Browser: Chrome’s Approach 14:50 - The Prompt API & Gemini Nano 17:34 - Browser Support & Model Limitations 19:14 - Specialized AI APIs: Translator, Language Detector, Summarizer 21:08 - Translator & Language Detector API Demos 23:20 - Building a Complete In-Browser Translation App 26:18 - Multimodal Prompt API: Audio & Offline Transcription 28:58 - How the Offline Translation Pipeline Works 30:00 - Other Use Cases & Privacy Benefits 32:13 - Getting Started & Next Steps 33:42 - Closing Remarks & Resources