Case Study - A social network built on voice

Overview

Text-based social feeds reward fast typists and quick takes. Chirp asks a different question: what does a social network look like when every post is your voice? It's a pixel-faithful reimagining of pre-X Twitter, with the same Twitter-blue palette, the same column layout, and the same feeds, replies, and DMs, except the unit of content is a short audio clip you record straight from the browser.

We built the platform end-to-end. Users record clips on the device, audio is uploaded to Google Cloud Storage with backend-signed playback URLs, and the feed renders as a continuous audio thread with a single global player. Claude Haiku reads each clip's transcript to suggest hashtags and categorize content. Likes, bookmarks, replies, and follows flow through an event-driven backend so feature work stays composable.

The product is mobile-first by design. Every layout was tuned on real iOS Safari with proper dynamic viewport handling, sticky title bars, and a single scroll wrapper so the feed feels native instead of webby. The result is a social product where the texture is human (voices, pauses, accents) instead of word count.

What we did