Request a source manifest
A manifest should connect delivered files with their source recording and the relevant show or session. Agree on stable identifiers before splitting long conversations into clips. This preserves the relationship between originals, edited audio, transcripts, and later annotations.
Specify editorial removal rules
Decide how advertising, introductions, music, third-party inserts, and production chatter are handled. State whether you need the full conversation or selected regions. If material is removed, ask how timestamps in associated transcripts are updated or mapped.
Agree on text and speaker conventions
Describe transcript style, timestamp detail, and speaker-label consistency. Overlap, interruptions, and non-speech sounds should have explicit conventions where they matter. Speaker identity labels should contain only the information permitted and needed for the project.
Review technical and commercial fit separately
Check decoding, duration, track layout, alignment, and duplicate sources on representative files. Separately establish the permitted uses and any participant or content exclusions. A technically ready collection still needs the appropriate project-specific permissions and agreed delivery terms.
Review checklist
- Languages, topics, and conversation styles
- Hours and preferred speaker diversity
- Mixed audio or independent tracks
- Transcript, diarization, and edit exclusions
- Training use and participant-permission requirements