โ All workers
Transcribe
BetaTurn any podcast episode into a searchable, clickable transcript.
PodcastStandard / Friendly / Market Collaborative
What it does
Takes an audio feed item and produces a time-stamped WebVTT transcript. Runs entirely in-browser via Whisper (WebGPU/WASM). Market Collaborative tier shares the result so subsequent listeners skip the compute entirely.
Example output
JSONFeed extension fields attached to each processed item.
{
"_transcribe": {
"status": "complete",
"format": "webvtt",
"word_count": 8420,
"duration_seconds": 2814,
"language": "en",
"model": "whisper-large-v3",
"confidence": 0.91
}
}Enables downstream workers
Once this worker runs, these workers can process the result without re-reading the source.
Show NotesTranslationAd MarkersSentimentRelated Episodes
Use cases
- Search through a podcast like a webpage
- Quote a specific passage with a timestamp link
- Accessibility โ deaf and hard-of-hearing listeners
- Foundation for downstream workers (show notes, translation, ad detection)