Parses the generated TTS audio file with music-metadata to get the real
duration before computing word timings, eliminating desync between text
overlay and voice playback.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Video duration forced to 25-30s (content 22-27s + 3s ending)
- Logo watermark (bottom-right) visible throughout content section
- Ending slide: logo centered + store name with spring animation (3s)
- Text overlay now shows only the active spoken word (true TikTok sync)
- Server fetches logo from Cloudinary config and store name from user's pages
- Player preview also shows logo/storeName via /api/cloudinary/config + /api/pages
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Convert relative music track URLs to absolute before passing to Remotion renderer
- Reverts Dockerfile from Debian-slim back to Alpine + gcompat/libc6-compat for lighter image
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
MichaelandClaude Sonnet 4.6