voice-transcriber

Original source / Raw SKILL.md

---
name: voice-transcriber
description: Voice note transcription and archival for OpenClaw agents. Powered by Deepgram Nova-3. Transcribes audio messages, saves both audio files and text transcripts. Perfect for voice-first AI workflows, founder journaling, and meeting notes.
homepage: https://www.agxntsix.ai
license: MIT
compatibility: curl, jq, Deepgram API key
metadata: {"openclaw": {"emoji": "\ud83c\udf99\ufe0f", "requires": {"env": ["DEEPGRAM_API_KEY"], "bins": ["curl", "jq"]}, "primaryEnv": "DEEPGRAM_API_KEY", "homepage": "https://www.agxntsix.ai"}}
---

# Voice Transcriber 🎙️

Audio transcription via Whisper + voice note journaling.

## When to Use

- Transcribing voice messages or audio files
- Saving voice notes with full transcripts
- Converting speech to text from any audio format

## Usage

### Transcribe audio
```bash
bash {baseDir}/scripts/transcribe.sh /path/to/audio.ogg
bash {baseDir}/scripts/transcribe.sh /path/to/audio.ogg --out /path/to/output.txt
```

### Save voice note with transcript
```bash
python3 {baseDir}/scripts/save_voice_note.py /path/to/audio.ogg "Optional context"
```

## Supported Formats

OGG, MP3, WAV, M4A, FLAC, WEBM

## Credits
Built by [M. Abidi](https://www.linkedin.com/in/mohammad-ali-abidi) | [agxntsix.ai](https://www.agxntsix.ai)
[YouTube](https://youtube.com/@aiwithabidi) | [GitHub](https://github.com/aiwithabidi)
Part of the **AgxntSix Skill Suite** for OpenClaw agents.

📅 **Need help setting up OpenClaw for your business?** [Book a free consultation](https://cal.com/agxntsix/abidi-openclaw)


---

## Skill Companion Files

> Additional files collected from the skill directory layout.

### _meta.json

```json
{
  "owner": "aiwithabidi",
  "slug": "voice-transcriber-pro",
  "displayName": "Voice Transcriber Pro",
  "latest": {
    "version": "1.0.0",
    "publishedAt": 1771143858793,
    "commit": "https://github.com/openclaw/skills/commit/f5087869ea2d6bc1743453b3a446e7df99305d6c"
  },
  "history": []
}

```

### scripts/save_voice_note.py

```python
#!/usr/bin/env python3
"""
Save a voice note with transcript to the founder journal.
Stores both the audio file reference and full transcript.

Usage:
  save_voice_note.py <audio_path> <transcript_text> [--date 2026-02-15]
"""
import argparse
import os
import shutil
from datetime import datetime

VOICE_DIR = os.path.expanduser("~/.openclaw/workspace/memory/voice-notes")
JOURNAL_DIR = os.path.expanduser("~/.openclaw/workspace/memory/founder-journal")

def save(audio_path: str, transcript: str, date: str = None):
    date = date or datetime.now().strftime("%Y-%m-%d")
    time_str = datetime.now().strftime("%H%M")
    
    # Copy audio to voice-notes archive
    ext = os.path.splitext(audio_path)[1] or ".ogg"
    audio_dest = os.path.join(VOICE_DIR, f"{date}_{time_str}{ext}")
    if os.path.exists(audio_path):
        shutil.copy2(audio_path, audio_dest)
        print(f"Audio saved: {audio_dest}")
    
    # Append to daily journal
    journal_path = os.path.join(JOURNAL_DIR, f"{date}.md")
    existed = os.path.exists(journal_path)
    
    with open(journal_path, "a") as f:
        if not existed:
            f.write(f"# Founder Journal — {date}\n\n")
        f.write(f"## Voice Note ({datetime.now().strftime('%I:%M %p')})\n")
        f.write(f"**Audio:** `{audio_dest}`\n\n")
        f.write(f"**Transcript:**\n{transcript}\n\n")
        f.write("---\n\n")
    
    print(f"Journal updated: {journal_path}")
    return journal_path

if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument("audio_path")
    parser.add_argument("transcript", nargs="?", default="[No transcript available]")
    parser.add_argument("--date")
    args = parser.parse_args()
    save(args.audio_path, args.transcript, args.date)

```

### scripts/transcribe.sh

```bash
#!/bin/bash
# Transcribe audio using OpenRouter's OpenAI-compatible API (whisper)
# Usage: transcribe.sh /path/to/audio.ogg [--out /path/to/output.txt]

AUDIO_FILE="$1"
OUT_FILE="${2:-}"
OR_KEY="${OPENROUTER_API_KEY:-$(grep OPENROUTER_API_KEY ~/.openclaw/workspace/.env 2>/dev/null | cut -d= -f2)}"

if [ -z "$AUDIO_FILE" ] || [ ! -f "$AUDIO_FILE" ]; then
    echo "Error: Audio file not found: $AUDIO_FILE" >&2
    exit 1
fi

# Use OpenAI directly for Whisper (OpenRouter doesn't support audio transcription endpoint)
# Fall back to local ffmpeg + sending to a model
TRANSCRIPT=$(curl -sf "https://api.openai.com/v1/audio/transcriptions" \
    -H "Authorization: Bearer ${OPENAI_API_KEY:-notset}" \
    -F "file=@$AUDIO_FILE" \
    -F "model=whisper-1" \
    -F "response_format=text" 2>/dev/null)

# If OpenAI key not available, try OpenRouter with a workaround
if [ -z "$TRANSCRIPT" ] && [ -n "$OR_KEY" ]; then
    # Convert to base64 and ask an LLM to transcribe via multimodal
    # For now, use ffmpeg to convert to wav and use a simpler approach
    TRANSCRIPT=$(curl -sf "https://openrouter.ai/api/v1/audio/transcriptions" \
        -H "Authorization: Bearer $OR_KEY" \
        -F "file=@$AUDIO_FILE" \
        -F "model=openai/whisper-large-v3" \
        -F "response_format=text" 2>/dev/null)
fi

if [ -z "$TRANSCRIPT" ]; then
    echo "[Audio transcription unavailable - no API key configured]"
    exit 0
fi

if [ -n "$OUT_FILE" ]; then
    echo "$TRANSCRIPT" > "$OUT_FILE"
fi

echo "$TRANSCRIPT"

```