| name | youtube-acquire |
| description | Acquire a YouTube video, playlist, channel, or transcript into the local yt-insights corpus. Use for requests such as recuperer une video, ajouter une chaine, download a transcript, or ingest a YouTube playlist. Do not use for code changes to a video player, YouTube SEO, or read-only corpus search. |
Acquire YouTube source material
Use the packaged yt-insights CLI. Keep acquisition in the main session because it can access the network and modify the local corpus.
Workflow
- Run
yt-insights doctor --json. Stop on failed required checks and report the failed check without exposing configuration values.
- Build the preview as
yt-insights acquire SOURCE --dry-run --json, replacing SOURCE with the user's source and adding only requested filters such as --slug, --years, --lang, or --analyze.
- Summarize the preview's selected count, source kind, output target, transcript language, and whether analysis is enabled. Report exclusions and discovery errors.
- If the source is one video and the user explicitly requested acquisition, repeat the command without
--dry-run.
- If the source is a channel, playlist, or batch, wait for explicit confirmation of the preview. Then repeat the same command without
--dry-run and with --yes.
- Report
selected, transcripts_ready, insights_ready, and the number and details of items that failed. Retain the preview's output target and report it with the final counts; do not invent more specific paths than the CLI returned.
Do not replace this workflow with a downloader command or direct corpus edits. Do not add --yes to the preview. If the CLI reports an authentication or bot-check failure for which browser cookies are relevant, suggest a new explicit preview using --cookies-from-browser chrome; do not add it preemptively.
Never claim that a transcript or insight is ready from the preview alone.
