elevenlabs-transcribe

To transcribe audio with speaker labels, run the built-in CLI script on your local file. The Skill uses ElevenLabs Scribe to provide speaker diarization, automatically labeling different speakers within the generated text transcript.

Yes, real-time speech-to-text streaming from a microphone is supported. By executing the transcription script with the microphone flag, the Skill captures live audio input and instantly converts it into continuous text captions.

Batch transcription requires Python 3.8+, ffmpeg, and a valid ELEVENLABS_API_KEY configured in your environment. These dependencies allow the script to process local audio files and generate structured JSON output with timestamps.

Yes, multilingual transcription works for streaming URLs. The Skill applies ElevenLabs Scribe to streaming audio sources, converting spoken language into searchable text while supporting multiple languages during the real-time transcription process.

The best way to generate searchable text from a conference recording is batch transcription. The Skill processes local audio files to create plain text or structured JSON with timestamps and optional speaker diarization for post-event notes.