Compress images for the web, in batches, on your Mac
Images are usually the heaviest thing on a web page and the easiest thing to fix. Smol batch-compresses your site images locally on a Mac — JPEG typically 60–85% smaller, with AVIF and WebP output — so pages load faster without a build step, an upload queue or a monthly bill.
Why image weight is usually the fastest performance win
On most sites, images account for the majority of transferred bytes. That makes them the dominant input to Largest Contentful Paint, which is the Core Web Vitals metric most sites fail. Shaving a 2 MB hero image down to 200 KB moves that number more than almost any amount of JavaScript tuning.
The reason it often goes unfixed is friction. Compressing images properly means either a build pipeline someone has to maintain, or a browser tool that caps free users and requires uploading client work to a third party. Neither is appealing when you just need this folder of photos to be smaller before you ship.
Batch processing that holds up at real volume
Drag in a folder of 500 mixed files and Smol processes all of them. You get per-file progress and per-file savings as it works, so you can see which assets were actually the problem. Virtual scrolling keeps the interface responsive past 1,000 items.
You can cancel mid-batch, and re-optimise specific files afterwards if a particular image needs a different setting. If compressing a file would make it larger, Smol keeps the original — so running a batch is never destructive to files that were already well optimised.
- No file size cap and no daily quota — the limit is your disk.
- Mixed batches: images, PDFs and video in the same run.
- Per-file savings reporting so you can spot the worst offenders.
- Strip EXIF, GPS and XMP metadata in the same pass.
Ship AVIF and WebP without a build pipeline
Smol writes AVIF and WebP directly. AVIF is typically 30–50% smaller than JPEG at equivalent quality with broad browser support, which for image-heavy pages is a substantial reduction in transferred bytes for a one-line change in your markup.
It also reads RAW — CR2, NEF, ARW, DNG — and HEIC, so photography can go from camera or phone to web-ready format without a detour through a separate editor.
Neural compression puts the bits where people look
A flat quality setting treats a face and a blurred background identically. Smol maps the image first — protecting faces, text, edges and fine texture, compressing smooth sky, walls and shadow much harder — then encodes with that map.
For real website photography this matters more than it sounds. Product shots, team photos and hero images all have a small region carrying the meaning and a large region that is mostly filler. Spending the bit budget accordingly is how you get files that are dramatically smaller and still look right at full width.
Local processing, which is also the fast option
Nothing is uploaded. There is no account, no server and no queue, which matters when the images belong to a client rather than to you.
It is also simply quicker. Online tools spend most of their wall-clock time moving your file across the network twice. Smol uses your Mac's CPU and GPU directly, so a batch that would take a browser tool most of an afternoon finishes while you are still looking at it.
Frequently asked questions
- How do I compress images for a website on a Mac?
- Drop your images or a folder into Smol, choose an output format — AVIF and WebP are the best choices for the web — and compress. JPEG images typically come out 60–85% smaller and PNG files 30–60% smaller losslessly. Everything runs locally, so there is no upload step and no limit on how many files you process.
- What image format should I use for a fast website?
- AVIF for photographic content where you want the smallest files, with broad modern browser support. WebP is a good fallback with slightly longer-standing compatibility. JPEG remains the safest universal option. Smol outputs all three, so you can test on your own images rather than guess.
- Can I compress hundreds of images at once?
- Yes. Drag a folder of 500 or more files — images, PDFs and videos can be mixed in the same batch — and Smol processes them all with per-file progress and savings. There is no batch size cap, and the interface stays responsive past 1,000 files.
- Will compressing images hurt their quality?
- Some loss is inherent to lossy compression, but Neural compression targets it. Smol protects faces, text, edges and fine detail while compressing low-detail regions harder, so the reduction lands where people are least likely to notice. You also keep direct control via the quality slider, and Smol keeps the original if compression would make a file larger.
- Does this replace an image build pipeline?
- For many sites, yes — particularly static sites, marketing pages and CMS-driven sites where images are uploaded by hand. Smol is a drag-and-drop step before upload rather than a build dependency to maintain. If you already have automated image processing in CI, Smol is more useful for one-off batches, client assets and RAW-to-web conversion.
- Is there a subscription or per-image cost?
- No. Smol is a one-time $29 purchase covering unlimited compression, including future updates. There are no per-file charges, monthly quotas or daily caps.