Home / MCP Tools / Voice Transcribe
Live MCP toolmedia

Voice Transcribe

Transcribe short audio to text via OpenAI Whisper.

Transcribe short audio to text via OpenAI Whisper. BYOK ONLY (your own OpenAI key). Pass 'audio_url' (fetched SSRF-safely, ≤10MB) or 'audio_b64'. Sync — chunk long recordings. Returns {ok, transcript, chars}.

Metered access — free-key quota or x402 pay-per-call (USDC on Base). Details at /pricing.

Parameters

ParameterTypeRequiredDescription
audio_urlstringoptionalhttp(s) URL of the audio (≤10MB)
audio_b64stringoptionalbase64 audio (alternative to audio_url)
content_typestringoptionale.g. audio/mpeg (helps format detection)
languagestringoptionalISO code hint, e.g. en (optional)
promptstringoptionalcontext/vocabulary hint (optional)
api_keystringoptionalyour OpenAI key (BYOK, inline)
key_refstringoptionalvault entry holding your OpenAI key (BYOK alt)
handlestringoptionalyour handle (for BYOK via vault)
secretstringoptionalyour agent secret (for BYOK via vault)

Call it over MCP (JSON-RPC 2.0)

curl -s https://wingmanprotocol.com/mcp \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -d '{"jsonrpc": "2.0", "id": 1, "method": "tools/call", "params": {"name": "voice_transcribe", "arguments": {"audio_url": "https://example.com", "audio_b64": "example", "content_type": "example"}}}'

Or plain HTTP

curl -s https://wingmanprotocol.com/tools/voice_transcribe \
  -H 'Content-Type: application/json' \
  -d '{"audio_url": "https://example.com", "audio_b64": "example", "content_type": "example"}'

Try it live

Edit the JSON and run it — this POSTs to https://wingmanprotocol.com/tools/voice_transcribe from your browser.

Related tools in Media

media
Gen Avatar Video
Async talking-avatar video via HeyGen (text→avatar).
media
Gen Id Portrait
FREE identity-preserving portrait: give ONE face photo (face_url or face_b64) and get a stylized portrait that keeps…
media
Gen Image
Generate an image from a text prompt via FAL Flux.
media
Gen Video
Async text-to-video via FAL (Kling/Veo).
media
Tts Voice
Text-to-speech via ElevenLabs.