Package Exports

@yemreak/stt
@yemreak/stt/index.js

This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (@yemreak/stt) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

Readme

@yemreak/stt

Usage

# Install
bun install -g @yemreak/stt

# Setup API key
stt config api_key sk_your_elevenlabs_key

# Transcribe audio
stt audio.mp3
→ Hello, this is the transcribed text

# From stdin
cat audio.mp3 | stt

# Word-level timestamps
stt --words meeting.mp4
→ 00:00.119-00:00.259	Hello	word	speaker_0
→ 00:00.280-00:00.520	world	word	speaker_0

# JSON output
stt --json podcast.mp3
→ {"text": "Hello world", "words": [...]}

What This Does

Speech to text with speaker diarization. Supports audio and video files.

Pattern:

Audio/video file → ElevenLabs API → transcript with timestamps and speakers
Multiple speakers detected automatically
Word-level timestamps for precise editing

Why it works:

ElevenLabs Scribe API (high accuracy, multi-language)
Auto speaker diarization (up to 8 speakers)
Video support via ffmpeg (extracts audio automatically)
Tab-separated output (pipe-friendly)

Example flow:

# Extract speaker 1's words
stt --words meeting.mp4 | grep speaker_1 | cut -f2

# Transcribe multiple files
fd -e mp3 . | xargs -I {} sh -c 'stt {} > {}.txt'

# Video to transcript
yt-dlp -x --audio-format mp3 $VIDEO_URL -o - | stt

# Word count by speaker
stt --words podcast.mp4 | awk -F'\t' '{print $4}' | sort | uniq -c

How to Build

# ElevenLabs Speech-to-Text API
curl -X POST https://api.elevenlabs.io/v1/speech-to-text \
  -H "xi-api-key: $API_KEY" \
  -F "file=@audio.mp3" \
  -F "model_id=scribe_v1" \
  -F "timestamps_granularity=word" \
  -F "diarize=true" \
  -F "num_speakers=8"

API: ElevenLabs Scribe v1

Config: ~/.config/stt/config.json or ELEVENLABS_API_KEY env var

License: Apache-2.0