Skill File

---
name: audio-transcriber
description: Transcribe audio files to text using AI Pass Whisper API.
version: 1.0.0
---

# Audio Transcriber Skill

Transcribe audio to text using Whisper.

## Requirements
- AI Pass API key (create at https://aipass.one/panel/developer.html)
- Set as environment variable: `$AIPASS_API_KEY`

## Usage

### Transcribe Audio (curl)
```bash
curl -s -X POST https://aipass.one/apikey/v1/audio/transcriptions \
  -H "Authorization: Bearer $AIPASS_API_KEY" \
  -F "model=whisper-1" \
  -F "file=@recording.mp3"
```

Response: `r.text` — transcribed text.

### With Language Hint
```bash
curl -s -X POST https://aipass.one/apikey/v1/audio/transcriptions \
  -H "Authorization: Bearer $AIPASS_API_KEY" \
  -F "model=whisper-1" \
  -F "file=@audio.wav" \
  -F "language=en"
```

### Python
```python
import requests, os

def transcribe(audio_path, language=None):
    files = {"file": open(audio_path, "rb")}
    data = {"model": "whisper-1"}
    if language:
        data["language"] = language
    
    r = requests.post(
        "https://aipass.one/apikey/v1/audio/transcriptions",
        headers={"Authorization": f"Bearer {os.environ['AIPASS_API_KEY']}"},
        files=files,
        data=data
    )
    return r.json()["text"]

text = transcribe("meeting.mp3")
print(text)
```

## Supported Formats
MP3, WAV, M4A, OGG, FLAC, WEBM, MPEG, MPGA

## Notes
- Model: `whisper-1` (OpenAI Whisper — high accuracy)
- Supports multiple languages and accents
- $1 free credit on signup at https://aipass.one
- Create API key at https://aipass.one/panel/developer.html

## Related
- App: https://aipass.one/apps/transcriber
- User guide: https://aipass.one/blog/ai-audio-transcriber-online
- Dev tutorial: https://aipass.one/blog/build-ai-transcriber