Minify images and graphics with one drop.

Not as of its latest release. Image Shrinker 1.6.4, from October 2020, ships x86_64 (Intel) compression binaries, so on an ARM-based Mac they only run through Rosetta 2. Shrinker Pro is a native ARM rewrite of it: the same drag-and-drop workflow and the same compressors (mozjpeg, pngquant, gifsicle and svgo), rebuilt from source as arm64 binaries.

Only through Rosetta 2. Its compression binaries are Intel code, so on an M1, M2, M3, M4, M5 or M6 Mac they run under translation, and fail outright on a Mac where Rosetta is not installed.

Because it contains Intel (x86_64) code. Rosetta 2 is the translation layer that lets Apple Silicon Macs run Intel apps, and macOS points you to it when you open one. Apple has published a sunset path for Rosetta 2: full availability through macOS 26 and 27, then only a limited subset for older games. An app that depends on it has a deadline. Shrinker Pro contains no Intel code, so it never needs Rosetta.

No. It is an independent rewrite. Image Shrinker is an Electron app by Stefan Schulz-Lauterbach, released under CC0-1.0. Shrinker Pro rebuilds the same idea in SwiftUI with the same underlying compressors (mozjpeg, pngquant, gifsicle and svgo), compiled as native arm64 binaries, and adds WebP, AVIF and HEIC support.

Shrinker Pro was built to be the native successor: the same drag-and-drop workflow, the same compressors, no Electron and no Rosetta. It is free, signed with a Developer ID and notarized by Apple.

Yes, on every Apple Silicon Mac: M1, M2, M3, M4, M5 and M6, and their Pro, Max and Ultra versions. It requires macOS 14 Sonoma or later, and there is no need to install Rosetta 2.

No, and that is deliberate. Apple ended Intel Mac support with macOS 26 Tahoe and Rosetta 2 is on a published sunset path, so the app is arm64-only by design. It requires macOS 14 Sonoma or later on Apple Silicon.

No. Compression happens on your Mac, using encoders bundled inside the app. Shrinker Pro makes no network requests except a daily update check, which sends no identifiers or usage data and never installs anything without asking you first.

Yes. No subscription, no account, no trial. It is distributed directly rather than through the Mac App Store, and its source is on GitHub under the MIT license.

In part. PNGs are compressed by pngquant 3.0.3, bundled as a native arm64 binary, so the app needs nothing installed alongside it. JPEGs go through mozjpeg, GIFs through gifsicle, SVGs through svgo, WebP through libwebp and JPEG XL through libjxl: one window, and the right encoder for each file. Since 1.2.0 there is also an optional command-line tool, shrinker, installed with Homebrew for scripts and build steps. It runs the same encoders, and the app still needs nothing.

For most formats, yes. That is where the savings come from. PNGs are reduced to an optimised palette by pngquant, JPEGs are re-encoded by mozjpeg at quality 75, and WebP and AVIF conversions encode at quality 80. Those are the Standard defaults: a Quality control in the session settings bar at the bottom of the window offers Super Low, Low, Standard or High for JPEG, WebP, JPEG XL, AVIF and HEIC, while PNG and GIF come out the same whichever you pick. Choosing one there applies until you quit; the default every session starts from is in Settings. GIF optimisation with gifsicle is lossless. Originals are kept by default, so you can compare the two before deciding which to keep.

Yes, in batches. HEIC and HEIF convert to JPEG by default: drop a folder of iPhone photos and each one comes out as an upright JPEG, with the original left where it was. You can choose WebP, AVIF or JPEG XL instead. Conversion rules are per input format, so converting HEIC never touches your PNGs.

Yes. Set the PNG or JPEG rule to WebP or AVIF in Settings and every file of that format you drop is converted, while other formats follow their own rules. For a one-off batch, “Convert all to” in the window’s session settings bar turns everything into JPEG, WebP, AVIF, JPEG XL or PNG for as long as the app is open.

Yes. Since 1.6.0, a JPEG converted to JPEG XL keeps the JPEG’s own compressed data rather than decoding and re-encoding it, so nothing is lost: the file comes out about a fifth smaller, and the original JPEG can be rebuilt from it byte for byte. Quality does not apply on that path. A JPEG that is rotated, resized or cropped on the way is encoded again at the chosen Quality instead. JPEG XL files can also be dropped in and shrunk like any other format.

Yes. Set a max size in the window and nothing comes out larger than that on its longest side. A 6000px photo capped at 2000px comes out 2000px wide, keeping its proportions, while anything already smaller is left at the size it arrived. It applies to the whole batch, GIFs included, and an animated GIF stays animated. SVG is vector, so it has no pixel size to cap.

Yes, from the center. Give it a ratio such as 16:9 and every image is cropped to that shape at whatever resolution it already had; give it a pixel size such as 1200x1200 and each one is cropped to that shape and then scaled to exactly that size, which is what you want when a folder of product photos all have to come out the same. Nothing is ever enlarged: an 800x600 image asked for 1200x1200 comes out 600x600, the right shape and smaller than asked. Animated GIFs keep every frame, and SVG is vector so it has no pixels to cut.

Yes. Drop a folder on the window and Shrinker Pro works through every supported image inside it, including subfolders. Hidden folders and packages such as apps and photo libraries are skipped, so a dropped folder never rewrites the inside of a Photos library. You can also select images in Finder and choose Open With → Shrinker Pro, or drop them on its Dock icon.

Because most compressors discard the EXIF orientation tag that tells an app which way up the photo goes. Shrinker Pro applies the rotation to the image itself instead of passing the tag along, so a portrait iPhone photo stays upright everywhere (in Preview, in Finder, and in a browser), whatever format it is converted to.

Yes. Metadata has three settings: keep everything, keep copyright and credit only, or keep nothing. The middle option keeps the fields that say who made the image and drops GPS coordinates, capture time and camera details, which is useful before posting a photo publicly.

Not unless you ask it to. By default the smaller copy is written beside the original with a .min suffix, so you end up with both. Replacing originals in place is a separate choice in Settings, and it is only offered when output is going into that same folder — send it to a folder you choose, or into a minified subfolder, and nothing of yours can be replaced at all. There the same choice simply decides whether the copy is called photo.min.png or photo.png.