transcribe

To transcribe audio with speaker diarization, you can use a CLI tool that processes single or multiple audio files and outputs diarized text. This Skill uses the bundled transcribe_diarize.py script to separate speakers and save results to the output directory.

Supported audio transcription output formats include plain text, JSON, and diarized JSON. You can configure these response formats via the CLI to suit different downstream processing needs for interviews or meetings.

Yes, you need an OpenAI API key to transcribe audio using this tool. The OPENAI_API_KEY environment variable is required for the bundled CLI to authenticate requests and generate transcripts.

Yes, you can transcribe multiple audio files at once. The CLI supports batch processing of audio files such as interviews, meetings, and lectures, saving all resulting transcripts under the output/transcribe directory by default.

To identify speakers in an audio transcript, you can use the diarization feature with known-speaker hints. This allows the transcription process to distinguish between different speakers and label their dialogue in the output.

Yes, you can specify optional language hints when transcribing audio. Providing a language hint helps improve transcription accuracy for recordings where the spoken language is known in advance.