Optimize images already in Studio
Open the file’s More actions menu and choose Optimize image. The worker uses the current presets to generate variants. The previous conversions remain available during processing and after a failed replacement. Open File information to compare each resulting file size with the original. The default keeps previously generated files. Choose Remove replaced conversions after the retention period to make old Studio-generated variants eligible for cleanup. The original file is preserved in both cases. For a large library, process images in bounded batches. The default command reports the selected batch. Add--apply to queue it. Use the printed cursor
with --after to resume. The principal must have write access to the workspace.
--refresh generates missing presets. Use --refresh to
rebuild existing Studio-owned variants, with --retention=keep by default.
Cleanup is opt-in and applies after processing.native.retirement_grace_seconds
(default 24 hours). Schedule the prune command if you want periodic cleanup.
For media managed by Spatie or an external provider, run that application’s own
conversion job and resync the source in Studio. The source retains ownership of
its files and its retention policy.
Generate video posters and playback previews
Video processing is optional and uses a background queue. SetFLEX_MEDIA_STUDIO_FFMPEG to the absolute path of the FFmpeg executable.
On macOS, Homebrew’s ffmpeg-full includes both poster encoders.
libwebp and the AVIF poster uses libaom-av1. The
optional MP4 playback preview uses libx264. Check the executable with
ffmpeg -encoders. Downloads retain the original. Files awaiting an MP4
preview show Processing video in the library.
Recover interrupted processing
Run the scheduler and register the bounded reconciliation command.Share identical small-image outputs
Native image batches process the largest requested bounds first. When smaller profiles resolve to the same geometry and compatible settings, their named renditions reference the already-published object rather than encoding and storing another copy. With the default profile, an 80×40 image has one physical AVIF at large quality 50 while thumbnail, medium and large remain available. Sharing is scoped to the source version and processing batch. Different formats, crop geometry, metadata settings or destination disks remain separate. A shared object is protected while any rendition or asset version still references it. Disable it globally:Verified AVIF output
AVIF requires encoder support in the installed GD or ImageMagick build. Automatic driver selection uses an available encoder; selecting an unsupported driver reports a conversion failure instead of saving another format with an.avif extension. Before publication, Studio checks the encoded file’s actual MIME type, dimensions and AVIF container brand.
Quality and effort settings control compression, but different encoders may produce different file sizes from Sharp. AVIF is not guaranteed to be smaller than WebP for every image.
Generate images without a permanent queue worker
Use command mode if your hosting can run Laravel scheduled commands but does not maintain a queue worker:--documents separately for document thumbnails; it cannot be combined with --images.
--limit bounds candidate rendition rows, not the number of output files: one admitted image batch can generate its configured variants together. The time budget stops admitting new batches; it does not interrupt an encoder already running. Keep source size, pixel and memory limits appropriate for your server. A failed image is recorded and does not prevent later admitted image batches from being attempted; the command returns a failure exit code if any batch fails. Existing attempt limits still apply.
The default image_execution is queue, preserving the configured queue connection’s behavior. Processing still runs on your PHP server and uses temporary local space even when originals and all generated files are stored on S3.