Give it a voice—in both directions. Free, built into the browser, no paid services:

— Speak every answer aloud using the Web Speech API (speechSynthesis). Prefer a British English voice if the system has one. — Add a 🎙 microphone button using webkitSpeechRecognition (Chrome): click, speak, and send the transcript through the same /chat flow. — Show a status line while listening and thinking ("● listening…", "● thinking…"). — Browsers block audio until the user interacts with the page — make sure speech starts reliably after the first click. — Give me a test plan: list what I should click and say to verify that both directions work.

Checkpoint: now you're talking to your notes — and they're answering.