Zum Inhalt springen
Kitbay
Skill

speech-to-text

MITAktualisiert am 9. Nov. 202679.567GitHub-Sterne

A copy, kept in the OpenMontage repository, of ElevenLabs' skill for transcribing audio and video with its Scribe v2 speech-to-text API.

Was der Skill vermittelt

A copy, kept in the OpenMontage repository, of ElevenLabs' skill for transcribing audio and video with its Scribe v2 speech-to-text API. It shows the agent how to call the API with cURL, the Python SDK or the JavaScript SDK, including word-level timestamps, speaker labels, up to 100 key terms to improve recognition, and language hints, plus a real-time model for live transcription through a WebSocket client or a React hook. It covers more than 90 languages, files up to 3 GB or 10 hours, common audio and video formats and the main error codes. It needs internet access and an ElevenLabs API key in the ELEVENLABS_API_KEY environment variable.

Funktionen

  • Transcribe audio or video files
  • Label speakers and add word timestamps
  • Stream live transcription
  • Boost recognition of key terms