Say something.

Click the button or tap the space bar. Click again to pause, double-click or hold to finish — then what you said becomes a draft you fix by talking to Frai, and send on.

Your words stay in this browser, and for the testing phase so does the sound of every capture (the newest 50). Nothing leaves until you send a draft to your own AI — and every new capture is reviewed by it once, shown beside your words. What leaves this device