Whisper Transcriber
@taylor/whisper-transcriberv2026.10.14.1
01README
Transcribe meeting audio locally with Whisper (faster-whisper, large-v3 by default) in a container under Apple's container CLI or Docker. Writes a timestamped text transcript plus SRT, WebVTT and JSON with word-level timestamps; the image builds itself on first use and model weights are cached between runs.
02Behaviors
@taylor/whisper-transcriberv2026.10.14.1whisper_transcriber.ts
Global Arguments
| Argument | Type | Description |
|---|---|---|
| runtime? | enum | Container runtime: Apple `container`, `docker`, or `auto` (prefers Apple container when running) |
| image? | string | Image repository name; the tag is derived from a hash of the embedded recipe |
| model? | string | Whisper model name (large-v3 is the most accurate) |
| computeType? | enum | CTranslate2 precision; float32 is the highest quality on CPU |
| cpus? | number | CPUs given to the container (default: up to 8, never more than the host has) |
| memory? | string | Memory given to the container (Apple container defaults to 1G, too small for large models) |
| cacheDir? | string | Host directory for cached model weights (default ~/.cache/swamp-whisper) |
fn build(noCache?: boolean)
Build the whisper container image
| Argument | Type | Description |
|---|---|---|
| noCache? | boolean | Rebuild without layer cache |
fn transcribe(audioPath: string, outputDir?: string, language?: string, beamSize?: number, vad?: boolean, initialPrompt?: string)
Transcribe an audio file to .txt/.srt/.vtt/.json (builds the image if missing)
| Argument | Type | Description |
|---|---|---|
| audioPath | string | Path to the audio file (relative paths resolve from the repo root) |
| outputDir? | string | Where to write .txt/.srt/.vtt/.json (default: next to the audio) |
| language? | string | ISO language code, e.g. en (default: auto-detect) |
| beamSize? | number | Beam search width |
| vad? | boolean | Skip silence with Silero VAD (reduces hallucinations) |
| initialPrompt? | string | Vocabulary hint: names, jargon, acronyms used in the meeting |
Resources
image(infinite)— The built whisper container image
transcript(infinite)— Metadata and output locations for a transcript
Files
text(text/plain)— Timestamped plain-text transcript
03Stats
A
93 / 100
Downloads
0
Archive size
28.9 KB
- Has README or module doc2/2earned
- README has a code example1/1earned
- README is substantive1/1earned
- Most symbols documented1/1earned
- No slow types (deprecated)1/1earned
- Dependencies pass trust audit2/2earned
- Has description1/1earned
- Platform support declared (or universal)2/2earned
- License1/2partial
no SPDX license declared
- Verified public repository2/2earned
Repository
https://github.com/taylorific/swampworksdeno-command1 accepted
Spawns a subprocess with Deno.Command().
- models/whisper_transcriber.ts:190 — runs the container or docker CLI; argument lists are built in this file, see README What it runs
05Platforms
06Labels