# Ticket: YouTube watch-history fallback sources

## Metadata
- Type: Ticket
- Status: Archived
- Project: YouTube-to-Obsidian Knowledge Capture
- Created: 2026-05-15
- Updated: 2026-05-17
- Priority: High

## Goal

Find a practical fallback for automated or low-friction watched-video import now that Google Data Portability is unavailable in the user's country/region.

## Why

Daily YouTube notes need actual watched-video identity and timestamps. Manual Google Takeout is not the desired workflow.

## Scope

Included:
- Investigate Android-side options for exporting/capturing YouTube history or current video context, accounting for possible sideload/F-Droid restrictions on the Android tablet.
- Investigate browser/session-based approaches only as a last resort and with explicit security review.
- Investigate Google My Activity email/notification/export alternatives.
- Investigate whether YouTube app sharing/current video URL can be captured with low friction.
- Investigate periodic user-assisted capture workflows that are less manual than Takeout.
- Document risks and choose a fallback strategy.

Not included:
- Scraping Google credentials/passwords.
- Bypassing YouTube app protections or DRM.
- Writing to Obsidian.

## Candidate Approaches

1. **Android automation / share intent**
   - Use Android share sheet or automation app to capture current YouTube URL when the user can tap Share.
   - Less ideal because user interaction is limited.

2. **Notification/accessibility/task automation**
   - Tools such as Tasker/MacroDroid/Automate might capture foreground app title or notification/media metadata.
   - Needs careful privacy/security review and may be brittle.
   - Prefer apps available through Google Play if F-Droid/sideloading is blocked on the tablet.

3. **Browser automation against My Activity**
   - A headless/browser session could read My Activity after user login.
   - High security and brittleness concerns; last resort only.

4. **Semi-automated My Activity export/review**
   - User periodically opens My Activity and saves/exports page data into Nextcloud.
   - Not ideal, but less heavy than full Takeout.

5. **Voice-first matching only**
   - Continue with voice captures and rely on title/channel/topic spoken by the user.
   - Use YouTube search/API metadata enrichment after a likely video is identified.

## Acceptance Criteria

This ticket is done when:
- [ ] At least three fallback options are assessed for feasibility, security, and user effort.
- [ ] A preferred fallback strategy is selected.
- [ ] Next implementation ticket is created.

## 2026-05-17 Spec Advancement

Confidence level: medium for candidate ordering; low for final approach until Android constraints are tested.

Decisions now stable:
- Avoid credential scraping and browser automation unless all lower-risk options fail.
- Prefer Google Play/stock Android-compatible approaches because F-Droid/sideloading may be blocked.
- Voice-first matching remains the safe baseline.

Refined next milestone:
- Test two low-risk fallbacks: manual/share URL capture and Android-accessible metadata capture, then compare against voice-only matching.

Updated next actions:
- [ ] Identify already-installed or Google Play-available automation/share tools.
- [ ] Test whether Share from YouTube can send URL/title into Nextcloud or a local text file with minimal friction.
- [ ] Test whether voice transcript plus approximate date can identify videos well enough using public metadata/search.
- [ ] Document security risk before any browser/session-based My Activity approach.

## Notes

- Google Data Portability is blocked because Google reports it is unavailable in the user's country/region.
- Manual Takeout remains a diagnostic fallback, not the desired operating model.

## 2026-05-15 Android App Distribution Note

The user reports that sideloading/third-party app stores such as F-Droid may not be available or may be blocked on the Android tablet. Fallback investigations should prioritize apps available through Google Play, already-installed apps, or workflows that do not require installing privileged tools.

## Archive Note

Archived because the YouTube-to-Obsidian project is dead in the water for now. Retained for future reference if requirements change.