Untethered

What if social media let you be honest?

Every social platform starts the same way — a profile, a photo, a performance. Untethered starts with a question: what would you say if nobody knew it was you?

This was my capstone project — rethink social media from scratch. Not a redesign. Not a clone. A completely different idea. Anonymous audio. No faces, no followers, no clout. Just a microphone and something to say.

You record “rants” — short voice posts — trim them down, tag them, and share them to a public feed. Others listen, like, and reply. No one knows who said it. All that matters is the thought.

Designed first, then built to match

Every screen was designed in Figma before writing a single line of code. The colors — sky blue and warm yellow — were picked to feel calm and human. Social media is already loud enough.

I mapped out every interaction before building it: recording a rant, editing it, publishing, browsing the feed, replying in threads. Each screen was designed as a prototype, then built to look exactly the same in the real app.

The goal wasn't making it look perfect — it was making it feel right. If a design looked good on paper but didn't work in the real app, the design changed — not the code.

Every animation has a job

When you open the app, a balloon animation plays while things load — built with Rive, a tool that makes animations lightweight and interactive without slowing the app down.

The little details matter most: your phone vibrates when you tap buttons, gives you a subtle buzz when something succeeds, and a gentle pulse when you scroll. The waveform moves while you record, reacting to your voice in real time so you know the mic is working.

Nothing moves just to look cool. The loading balloon tells you something is happening. The vibrations confirm your tap landed. The waveform proves the mic is listening. If an animation doesn't help you, it doesn't belong.

Designed to swap parts without breaking the whole

The app is built so every major piece — login, storage, audio, likes — is a separate module that plugs in like a Lego brick. If I ever need to swap out the database or switch cloud providers, I change one piece without touching the rest of the app.

Audio was the hardest part. You can pause mid-recording, pick back up, record multiple clips, then trim, cut, and stitch them together before posting. The finished file uploads to Google Cloud through a secure handoff so nothing gets lost or intercepted.

Every user's data is locked down — you can only access what's yours. Likes, comments, and threaded replies all work the way you'd expect. And future features like voice disguising, transcription, and moderation already have placeholder slots in the codebase, ready to be built.

What I'd build next

AI voice changing so people can speak freely without anyone recognizing their voice. Automatic text versionsof every rant so they're searchable and accessible to everyone. A smarter feedthat shows you what's resonating with people, not just what's newest.

The groundwork is already laid. These aren't ideas on a whiteboard — the app is built to plug them in.