wtf is speech.js?

Speech.js is a JavaScript library that lets you add voice recognition to web pages. Instead of typing or clicking, users can speak into their browser's microphone and have their words converted to text, useful for search boxes, voice commands, note-taking apps, or accessibility features. Under the hood, it's a simplified wrapper around Chrome's built-in speech recognition API. Rather than wrestling with the raw API's quirks, you create a Speech object with a few settings, then listen for events like "the user started speaking" or "here's what they said." The library handles the plumbing: it captures audio, sends it to Google's servers (which do the actual speech-to-text work), and sends results back to your code. You then decide what to do with the recognized words, display them on the page, submit a search, trigger an action, whatever you want. The library supports 80+ languages and regional accents, from English in Australia to Mandarin Chinese. You can also fine-tune behavior: enable continuous listening so it doesn't stop after one sentence, show interim results as the user is still speaking (so they see partial text updating in real time), or auto-restart if the connection drops. A debugging mode logs everything to the browser console to help you troubleshoot. Who would use this? Any web developer building voice-first features, a voice search bar, a hands-free note app, a voice-controlled game, or a smart home dashboard. It's particularly valuable for accessibility, letting users who can't type or prefer speaking interact with your site. Since it wraps a browser feature rather than building speech recognition from scratch, it's lightweight and relies on infrastructure that's already there in Chrome.