How do I install the Whatsapp Voice Transcriber skill?

Run npx skillmds@latest add rushindrasinha/whatsapp-voice-transcriber in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Whatsapp Voice Transcriber safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Whatsapp Voice Transcriber?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Whatsapp Voice Transcriber free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Whatsapp Voice Transcriber?

rushindrasinha (@rushindrasinha) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Whatsapp Voice Transcriber by rushindrasinha

Name: Whatsapp Voice Transcriber
Author: rushindrasinha

Whatsapp Voice Transcriber

Use when a user sends a voice note and wants it transcribed, or says "transcribe this voice note", "what did I say?", "convert voice to text", or "transcribe this audio". Handles OGG, MP3, WAV, and other audio formats. Outputs transcription with content classification (Idea/Task, Draft, or Note). Uses mlx_whisper large-v3-turbo for Apple Silicon — fully local, no API cost.

File contents

WhatsApp Voice Transcriber

Transcribes voice notes locally using mlx_whisper (Apple Silicon) and classifies by content type. No API cost, runs fully offline.

Requirements

mlx_whisper (recommended): pip install mlx-whisper — uses large-v3-turbo model on Apple Silicon
Fallback: openai-whisper (pip install openai-whisper) — base model
Python 3.10+
OpenClaw CLI (optional, for send-back delivery)

Usage

python3 scripts/transcribe.py <audio_file_path>

Example:

python3 scripts/transcribe.py /tmp/voice_note.ogg

Configuration (env vars)

Variable	Default	Description
`MLX_WHISPER_PATH`	`/opt/homebrew/bin/mlx_whisper`	Path to mlx_whisper binary
`WHISPER_MODEL`	`mlx-community/whisper-large-v3-turbo`	Whisper model to use
`NOTIFY_TARGET`	(empty)	Phone/JID to send result to via openclaw
`NOTIFY_CHANNEL`	`whatsapp`	Channel for notification
`OPENCLAW_PATH`	`/opt/homebrew/bin/openclaw`	Path to openclaw binary

Set NOTIFY_TARGET to auto-send transcription after processing:

NOTIFY_TARGET="+1234567890" python3 scripts/transcribe.py /tmp/voice.ogg

Classification

Heuristic classifier labels each transcription:

💡 Idea / Task — keywords: remind, task, build, idea, need to
✉️ Draft — keywords: tell, message, send, write, reply
📝 Note — everything else

OpenClaw Integration

Wire into your agent skill dispatcher to auto-transcribe incoming voice notes:

# In your OpenClaw agent logic:
import subprocess, tempfile, os

def handle_voice_note(attachment_path):
    result = subprocess.run(
        ["python3", "~/clawd/.agents/skills/whatsapp-voice-transcriber/scripts/transcribe.py",
         attachment_path],
        capture_output=True, text=True
    )
    return result.stdout.strip()

Model Notes

mlx_whisper large-v3-turbo — best for Apple Silicon (M1/M2/M3). Highly accurate, fast enough for real-time. ~1GB model download on first use.
whisper base — fallback only. Much lower accuracy. Use if no Apple Silicon.
Model is downloaded automatically on first run.

Scripts

scripts/transcribe.py — Main script. Transcribes audio, classifies, prints, and optionally sends.

---name: whatsapp-voice-transcriberdescription: Use when a user sends a voice note and wants it transcribed, or says "transcribe this voice note", "what did I say?", "convert voice to text", or "transcribe this audio". Handles OGG, MP3, WAV, and other audio formats. Outputs transcription with content classification (Idea/Task, Draft, or Note). Uses mlx_whisper large-v3-turbo for Apple Silicon — fully local, no API cost.---# WhatsApp Voice TranscriberTranscribes voice notes locally using mlx_whisper (Apple Silicon) and classifies by content type. No API cost, runs fully offline.## Requirements- **mlx_whisper** (recommended): `pip install mlx-whisper` — uses `large-v3-turbo` model on Apple Silicon- **Fallback**: openai-whisper (`pip install openai-whisper`) — base model- Python 3.10+- OpenClaw CLI (optional, for send-back delivery)## Usage```bashpython3 scripts/transcribe.py <audio_file_path>```Example:```bashpython3 scripts/transcribe.py /tmp/voice_note.ogg```## Configuration (env vars)| Variable | Default | Description ||---|---|---|| `MLX_WHISPER_PATH` | `/opt/homebrew/bin/mlx_whisper` | Path to mlx_whisper binary || `WHISPER_MODEL` | `mlx-community/whisper-large-v3-turbo` | Whisper model to use || `NOTIFY_TARGET` | _(empty)_ | Phone/JID to send result to via openclaw || `NOTIFY_CHANNEL` | `whatsapp` | Channel for notification || `OPENCLAW_PATH` | `/opt/homebrew/bin/openclaw` | Path to openclaw binary |Set `NOTIFY_TARGET` to auto-send transcription after processing:```bashNOTIFY_TARGET="+1234567890" python3 scripts/transcribe.py /tmp/voice.ogg```## ClassificationHeuristic classifier labels each transcription:- 💡 **Idea / Task** — keywords: remind, task, build, idea, need to- ✉️ **Draft** — keywords: tell, message, send, write, reply- 📝 **Note** — everything else## OpenClaw IntegrationWire into your agent skill dispatcher to auto-transcribe incoming voice notes:```python# In your OpenClaw agent logic:import subprocess, tempfile, osdef handle_voice_note(attachment_path): result = subprocess.run( ["python3", "~/clawd/.agents/skills/whatsapp-voice-transcriber/scripts/transcribe.py", attachment_path], capture_output=True, text=True ) return result.stdout.strip()```## Model Notes- **mlx_whisper large-v3-turbo** — best for Apple Silicon (M1/M2/M3). Highly accurate, fast enough for real-time. ~1GB model download on first use.- **whisper base** — fallback only. Much lower accuracy. Use if no Apple Silicon.- Model is downloaded automatically on first run.## Scripts- `scripts/transcribe.py` — Main script. Transcribes audio, classifies, prints, and optionally sends.

Frequently asked questions