openai-whisper-api
openclaw/openclaw
Transcribe audio to text via OpenAI's Whisper API with gpt-4o, mini, diarization, or whisper-1 models.
What is openai-whisper-api?
Transcribe audio files to text using OpenAI's speech-to-text API. Supports multiple models including gpt-4o-transcribe, gpt-4o-mini-transcribe, speaker diarization, and the original whisper-1. Use this when you need to convert audio recordings into searchable, editable text.
- Transcribe audio files (mp3, mp4, m4a, wav, webm, ogg, mpeg, mpga) to text
- Support multiple transcription models: gpt-4o-transcribe, gpt-4o-mini-transcribe, gpt-4o-transcribe-diarize, and whisper-1
- Output transcriptions as plain text or JSON format
- Add speaker labels via diarization mode
- Specify language and provide context prompts to improve accuracy
- Use custom OpenAI-compatible proxies or local gateways via OPENAI_BASE_URL
How to install openai-whisper-api
npx skills add https://github.com/openclaw/openclaw --skill openai-whisper-api- curl and node installed (curl can be installed via brew)
- OPENAI_API_KEY environment variable set or configured in OpenClaw config
- Audio file in supported format (mp3, mp4, m4a, wav, webm, ogg, mpeg, mpga)
- File size under 25 MB for hosted OpenAI API
How to use openai-whisper-api
- 1.Set OPENAI_API_KEY environment variable or configure it in ~/.openclaw/openclaw.json
- 2.Run the transcribe script with your audio file: {baseDir}/scripts/transcribe.sh /path/to/audio.m4a
- 3.Optionally specify a model with --model (default is gpt-4o-transcribe)
- 4.Optionally specify output format with --out /path/to/output.txt or --json for JSON output
- 5.For speaker identification, use --model gpt-4o-transcribe-diarize
- 6.For language-specific transcription, add --language en (or other language code)
Use cases
- Transcribe meeting recordings or interviews for documentation and searchability
- Convert podcast episodes or video audio to text for accessibility and archival
- Identify and label speakers in multi-person conversations using diarization
- Batch process audio files with language-specific transcription
- Generate transcripts with custom context prompts for domain-specific terminology
- Developers integrating speech-to-text into applications
- Researchers and analysts processing audio data
- Content creators and podcasters needing transcripts
- Teams managing meeting recordings and documentation
openai-whisper-api FAQ
Supported formats include mp3, mp4, mpeg, mpga, m4a, wav, and webm. Maximum file size is 25 MB on the hosted API.
Use the --model gpt-4o-transcribe-diarize flag. This model automatically identifies and labels different speakers in the audio.
Yes, set the OPENAI_BASE_URL environment variable to point to an OpenAI-compatible proxy or local gateway.
gpt-4o-transcribe is the newer, more accurate model. whisper-1 is the original Whisper model. gpt-4o-mini-transcribe is a faster, lighter variant.
Yes, use the --prompt flag to provide speaker names or domain-specific context, except when using diarization mode which rejects prompts.
Full instructions (SKILL.md)
Source of truth, from openclaw/openclaw.
name: openai-whisper-api description: "OpenAI Audio Transcriptions API via curl; gpt-4o-transcribe, mini, diarize, or whisper-1." homepage: https://platform.openai.com/docs/guides/speech-to-text metadata: { "openclaw": { "emoji": "🌐", "requires": { "bins": ["curl", "node"], "env": ["OPENAI_API_KEY"] }, "primaryEnv": "OPENAI_API_KEY", "install": [ { "id": "brew", "kind": "brew", "formula": "curl", "bins": ["curl"], "label": "Install curl (brew)", }, ], }, }
OpenAI transcriptions API
Transcribe audio through /v1/audio/transcriptions. Set OPENAI_BASE_URL for an OpenAI-compatible proxy or local gateway.
Quick start
{baseDir}/scripts/transcribe.sh /path/to/audio.m4a
Defaults:
- Model:
gpt-4o-transcribe - Output:
<input>.txt
Useful flags
{baseDir}/scripts/transcribe.sh /path/to/audio.ogg --model gpt-4o-transcribe --out /tmp/transcript.txt
{baseDir}/scripts/transcribe.sh /path/to/audio.ogg --model gpt-4o-mini-transcribe
{baseDir}/scripts/transcribe.sh /path/to/audio.ogg --model gpt-4o-transcribe-diarize --json
{baseDir}/scripts/transcribe.sh /path/to/audio.ogg --model whisper-1
{baseDir}/scripts/transcribe.sh /path/to/audio.m4a --language en
{baseDir}/scripts/transcribe.sh /path/to/audio.m4a --prompt "Speaker names: Peter, Daniel"
{baseDir}/scripts/transcribe.sh /path/to/audio.m4a --json --out /tmp/transcript.json
Notes:
- Supported upload formats include
mp3,mp4,mpeg,mpga,m4a,wav,webm. - 25 MB upload limit on the hosted API.
- Use diarize for speaker labels; script sends
chunking_strategy=autoand rejects--prompt.
API key
Set OPENAI_API_KEY, or configure it in the active OpenClaw config file ($OPENCLAW_CONFIG_PATH, default ~/.openclaw/openclaw.json). Optionally set OPENAI_BASE_URL:
{
skills: {
"openai-whisper-api": {
apiKey: "OPENAI_KEY_HERE",
},
},
}
Related skills
More from openclaw/openclaw and the wider catalog.
