Steps
- Install peepshow
npm install -g peepshow - Run on the APNG
Format auto-detected from extension + content.
peepshow ./tutorial.apng - Feed frames to your LLM
Drag the frame folder into Claude / ChatGPT or pipe the JSON to an API.
peepshow ./tutorial.apng --emit json | jq '.frames[].path'
Why it works
APNG is the modern animated image format — losslessly compressed, supports transparency, widely used in tutorial assets and animated UI documentation. LLM vision APIs don't read it as motion. peepshow's format probe extracts every frame, drops near-duplicates, and emits a JPEG sequence.
When it helps
- Documentation hosted as APNG (technical tutorials, UI walkthroughs).
- Animated logos / loading states you want analysed.
- Asset archives where the source is APNG and conversion to MP4 isn't desired.
Pitfalls
- Transparency channel is dropped on JPEG export — pass `--format png` if you need it.
- Some APNG encoders use non-standard chunks; peepshow falls back to ffmpeg's apng decoder.