HALOWERK medienwerk Speaker Time

Choose this endpoint when you need speaker-level segmentation of a multi-participant audio or video file with precise timestamps per turn. It is ideal for meeting transcripts, podcast editing, legal recordings, or call center analytics where knowing who spoke when matters more than (or in addition to) what was said. Prefer this over a pure transcription endpoint when speaker attribution is the primary goal.

• File URL is inaccessible or behind authentication — returns an error indicating the resource could not be fetched
• File exceeds size or duration limits — returns an error with size/duration constraints
• Unsupported audio/video codec or container format — returns a format error
• No speech detected in the file — returns an empty or minimal speaker turn list
• Base64 payload is malformed or corrupted — returns a decode error
• Network timeout fetching a remote URL — returns a timeout error