Batch compress images to 100KB

Use cases

• Prepare portal or form uploads with a stated byte limit
• Reduce email attachment size across several photographs
• Create lightweight product thumbnails while keeping consistent dimensions

What the target-size search actually does

The target is an upper bound, not an instruction to produce exactly 100.0KB. If several tested qualities fit, the highest-quality candidate wins. The final file may therefore be noticeably smaller than the limit.

The search currently applies to JPEG and WebP. PNG uses lossless encoding plus OxiPNG optimization, while AVIF follows its format encoder settings rather than this target-byte search.

When 100KB is unrealistic

Noise, hair, foliage, text overlays and fine product texture are expensive to encode. A 4000-pixel image can remain too large even at low quality, while a simple 1200-pixel image may fit easily.

• Reduce maximum width before forcing quality lower.
• Compare JPEG and WebP instead of assuming one is always smaller.
• Inspect edges, text and gradients after aggressive compression.

Measured data instead of a universal claim

The linked BatchMint benchmark publishes source files, output files, exact codec versions and byte counts. It is reproducible, but its synthetic fixtures are not a promise about your photographs.

How local browser processing works

For each image, the worker encodes quality candidates between 40 and 92 using a bounded seven-step binary search. It keeps the highest tested quality whose encoded byte count is within the requested target. Candidate image bytes stay inside the local worker and are discarded after the final Blob is selected.

Selected images, filenames and watermark text are not sent to a BatchMint processing server. The browser still requests the site code, styles and codec assets used on demand.

How to use it

1. Choose the source images and select JPEG or WebP output.
2. Enter 100 in Target file size and optionally set a smaller maximum width.
3. Process locally and check the actual byte count and visual result for every file.

Sources and verification dates

Platform rules can change. Marketplace requirements below link to official pages, with the retrieval date and scope recorded for each source.

• BatchMint reproducible image-compression benchmark

  Deterministic fixtures, codec versions, settings and downloadable outputs generated July 17, 2026.

• BatchMint compression methodology

  Documented fixed pipeline and target-size algorithm.

Limitations

• A target cannot be guaranteed for every source image.
• PNG and AVIF do not currently use the target-size search.
• The search tests a bounded quality range; it does not endlessly re-encode.

FAQ