Current state
| Stage | Data | Where | Status |
|---|---|---|---|
| Synthetic demonstration | Fictitious scenario text only | Local files | Implemented; zero external requests |
| Narrow technical pilot | Up to five public search.list video records: IDs, title, channel ID/name, publication date and source URL | Local pilot artifact | One completed pilot; no comments, no pagination, no derived metrics |
| Full source discovery | Planned metadata-search output | Not written for live use | Blocked while unresolved |
Proposed workflow — not active
Research brief → official YouTube Data API → source selection → local minimum data store → independently labelled analysis → traceable report.
| Category | Purpose | Not used for |
|---|---|---|
| Video/channel IDs, titles, descriptions and dates | Traceability and source assessment | Identity profiling or market-size claims |
| Thread/comment IDs, public comment/reply text and dates | Only if permitted and implemented: analysis of expressed themes | Contacting authors, training a model, publishing full copies |
| Independently generated categories or metrics | Only if permitted: clearly labelled analytical output | Claiming the output is a YouTube metric or replacing YouTube data |
Controls not yet implemented
- Comment and reply collection through
commentThreads.list/comments.list. - Pagination, resume points and deduplication for comments.
- Automatic minimisation of author data.
- Automated 30-day refresh or deletion for all raw API data.
- Live analytical reporting and any model-assisted analysis of API data.
These controls are listed to avoid overclaiming. Their absence means the corresponding processing is not active.