nospeaky.ai

On your phone · nothing uploaded

Subtitles without
sending the video away.

Pick a clip. Your device runs the speech model and writes captions. NoSpeaky never receives your file. No home server. No cloud GPU bill for us.

Open Watch How it works

Best first test: 30–90 second video. First run downloads a small model, then it’s cached.

How it works

1. You pick a file Video or audio from your phone/camera roll.
2. Your device listens Whisper runs in the browser (WebGPU when available).
3. You watch + download Captions on screen, plus a normal .srt file.

Pricing (later)

Now

Free

On-device engine
Tiny/base models
SRT download

Later

Pro

~$5–8/mo

Longer clips helpers
Better models
URL tools if we add a safe proxy

Money features come after the on-device path feels good on real phones.

Status

live Site + HTTPS
live On-device Watch (Transformers.js Whisper)
later URL paste without a server

Built by Pi Studios. Your home network is not involved.