← 2026

Cerebro

A landing page that lip-syncs.

The landing page for Cerebro, a voice-capture device that sits on your desk and remembers what you tell it. The little face on the page mouths every phrase it shows, with no audio and no animation library.

GitHub
  • Next.js
  • TypeScript
  • Inline SVG
  • Web Animations API
  • CSS Modules

What it does

Mar 2026
  1. 01
    A face that talks
    The device on the page mouths short phrases one shape at a time, and each word appears as it's said.
  2. 02
    Kit on a desk
    The round AMOLED face sits on its stand in a drawn desk scene, with a shadow that matches its shape.
  3. 03
    Two voices
    Kit searches your memories and health data before it answers. Aly skips all that for quick back-and-forth.
  4. 04
    Inside the app
    Screens from the phone app: stress, offload and bandwidth scores, a health view, and pairing the device.
  5. 05
    The app's API lives here too
    Next.js API routes handle transcription, replies, speech and saved memories for the phone app.

Screens

Signature

The lip-synced face

The face on the page reads its phrases out with its mouth, not with sound. Each letter maps to one of six mouth shapes, and the mouth morphs from one to the next about every 110 ms while the words appear beside it. No sprite sheet, no video, no tweening library.

How it works

Design

DM Sans
Aa
  • #0D0D0D
  • #E8000A
  • #6B6A68
  • #F8F6F1