Stitch Captions

Whisper transcription · Ken Burns canvas engine · WebM export

Cloudflare Pages Function ready

1 Source Media

Drop one audio track and the still images that will carry it.

Choose audio file No track loaded
Choose images No images loaded

2 Output Format

The canvas viewport updates instantly when you switch orientation.

3 Motion & Transitions

How far each image drifts, and how one image hands over to the next.

4 Caption Engine

Word-level timings drive every animation on the 2D context.

5 Pipeline

Transcribe once, then preview or render as many times as you like.

Idle — load an audio track and images to begin.
Duration — Images 0 Words 0 Language auto

▶ Live Canvas Pipeline

Every frame is drawn in real time against the audio playback cursor.

RECORDING
Working…
Please keep this tab in the foreground.
0:00.0 / 0:00.0

⬇ Rendered Output

VP9 + Opus WebM at 5 Mbps, muxed straight from the canvas stream.

No render yet — your finished .webm will appear here.