Browser-only image processing. On browsers without native WebP encoding, third-party fallback codecs may be downloaded from CDNs; photographs are not uploaded to the CDN, Admin API, or R2.
Ready
None
JPEG EXIF capture date/year/decade extraction is enabled; HEIC metadata is experimental. Duplicate detection is for byte-identical files, not visually similar images. HEIC decoding attempts native support, then a CDN fallback. HEIC EXIF dates are best-effort; some files will still have no detected date. PNG/WebP metadata dates are not extracted. Original orientation is applied by the browser decoder. Legacy SHA-1 is provided for integration, but matching existing gallery duplicates requires checking the existing processor's digest semantics in Phase 3. ZIPs are read into memory once; extracted entries are processed individually. All output WebP blobs remain in memory for download until the next batch or page close; mobile browsers may still impose memory limits. *HEIC support depends on the device/browser.