About ez-stt
ez-stt is a local speech-to-text skill that runs ONNX Runtime with int8 quantization. It offers Parakeet models for English or multilingual transcription and Whisper for multilingual transcription, and requires ffmpeg.
Key features
- 01Local ONNX Runtime transcription
- 02Parakeet and Whisper backends
- 03Parakeet v2 and v3 model options
- 04Whisper Base for 99 languages
- 05Quiet mode and Matrix room option
Use cases
- 01Transcribing local audio files
- 02Transcribing English with Parakeet v2
- 03Choosing multilingual transcription with Parakeet v3 or Whisper
- 04Sending transcription to a Matrix room
Install
Copy the command and run it in a terminal.
$
npx skills add https:/After installing, type/ez-sttin Claude Code to confirm it loaded.
Bundled filesreferences/OPENCLAW.mdscripts/stt.py
Details
From the project READMEez-stt - Local Speech-to-Text
Unified local speech-to-text using ONNX Runtime with int8 quantization. Choose your backend:
- Parakeet (default): Best accuracy for English, correctly captures names and filler words
- Whisper: Fastest inference, supports 99 languages
Requires ffmpeg installed.
Usage
# Default: Parakeet v2 (best English accuracy)
scripts/stt.py audio.ogg
# Explicit backend selection
scripts/stt.py audio.ogg -b whisper
scripts/stt.py audio.ogg -b parakeet -m v3
# Quiet mode (suppress progress)
scripts/stt.py audio.ogg --quiet
Options
-b/--backend:parakeet(default),whisper-m/--model: Model variant (see below)--no-int8: Disable int8 quantization-q/--quiet: Suppress progress--room-id: Matrix room ID for direct message
Models
Parakeet (default backend)
| Model | Description |
|---|---|
| v2 (default) | English only, best accuracy |
| v3 | Multilingual |
Whisper
| Model | Description |
|---|---|
| base (default) | Fast multilingual (only Whisper variant in onnx-asr) |
Benchmark (24s audio)
| Backend/Model | Time | RTF | Notes |
|---|---|---|---|
| Whisper Base int8 | 0.43s | 0.018x | Fastest |
| Parakeet v2 int8 | 0.60s | 0.025x | Best accuracy |
| Parakeet v3 int8 | 0.63s | 0.026x | Multilingual |
OpenClaw
See OPENCLAW.md for OpenClaw-specific setup and openclaw.json configuration.
Comments
Markdown supported