Side-by-side
| peepshow | whisper.cpp / Whisper API standalone | |
|---|---|---|
| Audio transcription | ✅ via whisper.cpp + 4 cloud providers | ✅ whisper.cpp directly |
| Frame extraction | ✅ scene-detect, fps, dedup | ❌ transcription only |
| Sink fan-out | ✅ 95 destinations | ❌ |
| Run history / report | ✅ per-run report.html | ❌ |
| Setup | `npm i -g peepshow` + whisper.cpp | whisper.cpp alone |
| Cost | Same — whisper.cpp is local + free | Same |
| Multi-output emit | ✅ JSON / markdown / caveman | Plain text only |
Pick peepshow when…
- You also need frame extraction.
- Output needs to land in Notion / Slack / SQL / etc.
- You want a per-run report.html + run history.
- Multiple transcription providers (cloud + local) need to be one flag away.
Pick whisper.cpp / Whisper API standalone when…
- Audio-only — no frames needed.
- Streaming transcription on a live mic.
- Already heavily scripted around whisper.cpp's exact output.
Verdict
Use whisper.cpp directly for audio-only workflows. Use peepshow when frames + transcript + sinks are part of the same flow.