Batch image optimizer with a file size target
Add a batch of screenshots and product images, set a file size target in KB and get smaller files back, with a clear table of what changed. You can compare every result with the original at 100 percent and download a ZIP with a CSV report. It is free, needs no account and runs in your browser.
Heavy images slow down a product page and often break upload limits. Shrinking them by hand one at a time is slow, and guessing a quality setting rarely lands near the size you need. This tool searches for the quality for you, tells you honestly when a target cannot be reached, and lets you look at the result before you use it.
Optimize your images
Free, no login, and everything happens inside this tab. Your images are never uploaded, never stored and your originals are never changed. Files and results disappear when you reload the page.
Settings
How to use the image optimizer
- Drop up to 30 PNG, JPEG or WebP images on the drop area, or press Choose images. Each file can be up to 20 MB and 50 megapixels. A file that breaks a limit is skipped with a message that names it.
- Choose an output format. Auto encodes every format your browser can produce and keeps the smallest result. Options your browser cannot encode are disabled, so nothing is guessed.
- Optionally set a longest side in pixels and a target size in KB. With a target, JPEG and WebP get the highest quality that still fits. If even the lowest quality is too large, the image is scaled down step by step, but never below the smallest size you allow.
- Press Optimize. Images are processed one at a time, the progress is announced as text such as 3 of 10 done, and Cancel stops after the current image.
- Read the table. Each row shows dimensions and bytes before and after, the change, the format and quality used, a status and whether the target was met, with the reason when it was not. Press Compare to inspect an image with a slider or side by side at 100 percent.
- Download single files, or the ZIP with every result and the CSV report. The tool never changes or uploads your originals.
Example with made up numbers
Imagine a fictional dashboard screenshot saved as a PNG of 2400 by 1350 pixels and 1.8 MB. You set the format to WebP and the target to 300 KB. The tool tries the highest quality first, then the lowest, then narrows down in between, and keeps the highest quality that fits under 300 KB. If even the lowest quality were still too big, it would try 90, 80 and so on of the side length, stopping at your smallest allowed size, and the table would show the final size and whether the target was met. These numbers are invented, so your own results will differ.
Practical tips for screenshots and marketing images
Pick the format by content. Screenshots with sharp text and flat colour usually behave best as PNG or lossless looking WebP, while photos and gradients compress well as JPEG or lossy WebP. Lossy formats can leave soft halos around small text, so open Compare at 100 percent and read the smallest text before you accept a result.
Resize before you squeeze quality. An image that is wider than the space it is shown in wastes bytes, so set the longest side to what the page really needs. Keep room for sharper displays, which may need more pixels than the layout width. Scaling down is also the only way to shrink a PNG in the browser.
Compression can ruin an image when the target is too tight. Watch for blocky areas, smeared text and banding in gradients, and raise the target or the smallest allowed size instead of accepting them. When the table says the target was not met, that is information, not a failure to fix by lowering quality further.
Frame a screenshot first with the screenshot mockup exporter, then optimize the exported files here. For app icons and the small images browsers need, use the favicon package builder. For a full list of launch steps, see the product launch checklist.
Honest limitations
Encoding is done by your browser, so quality numbers are not comparable between browsers and the same settings can give different sizes. Some browsers cannot encode WebP, and very large images can fail on devices with little memory, in which case the row shows an error. There is no AVIF in this version. Quality here is a search aid, not a promise of lossless output, and PNG in the browser is only resized, never tuned. Files are handled one at a time to keep memory use low. Results are not saved, so download them before you leave the page.
Frequently asked questions
Are my images uploaded anywhere?
No. Images are read and encoded by your own browser and stay in the memory of this tab. Nothing is sent to a server and nothing is saved in your browser storage. Previews use temporary addresses that are released when you close them. Reloading the page removes every file and result, so download what you need first.
Why did the tool not reach my target size?
The table says why. For JPEG and WebP the tool first lowers quality, then scales the image down, but it stops at the smallest size you allowed so text does not become unreadable. PNG has no quality setting in the browser, so only scaling can shrink it. Choose WebP or JPEG, lower the smallest allowed size or raise the target.
What happens to transparency?
WebP and PNG keep transparent pixels. JPEG cannot store transparency, so transparent areas are filled with the background colour you choose, white by default. In Auto mode, JPEG is not offered for an image that has transparent pixels.
What does the CSV report contain, and is it safe to open in a spreadsheet?
One row per file with original_name, output_name, original_bytes, output_bytes, original_width, original_height, output_width, output_height, format, quality, target_bytes and target_met. Fields are quoted as described in RFC 4180. A text field that starts with an equals sign, plus, minus, at sign, tab or carriage return gets an apostrophe in front, so a spreadsheet shows it as text instead of running it as a formula.
Why is there no AVIF, and why do results differ between browsers?
This version covers WebP, JPEG and PNG. Encoding is done by the browser, so the same quality number can give a different file size in another browser or a newer version, and a browser may not support every format. The tool tests what your browser can encode and shows only those options.
Does optimizing remove metadata such as camera data or colour profiles?
Images are decoded and drawn again on a canvas before they are encoded, which usually does not carry over metadata such as camera data or an embedded colour profile. Colours are drawn as the browser shows them. If you rely on metadata or on exact colour management, keep the originals, which this tool never changes.
Ready to share your product?
Now that your screenshots are light and sharp, add them when you submit your product to RadarMakers.
RadarMakers is a free directory for indie projects and SaaS. Prepare a clear tagline, explain who your product helps and include real screenshots when you submit.