Compress an Image to an Exact Size in KB

Compress an image to an exact file size in KB

Most compressors give you a quality slider and leave you to guess. This one works the other way round: you name the file size you need, 20 KB, 50 KB, 100 KB, 200 KB, or anything in between, and it finds the highest quality that fits under that number. That is the version of the problem people actually have, because upload forms do not check quality, they check bytes.

It also resizes. If your target cannot be reached by lowering quality alone, the tool steps the pixel dimensions down and searches again, so you always end up with the largest, sharpest image that still passes the limit.

Free, unlimited, no signup, no watermark, and batch friendly. Nothing you drop here is uploaded.

How to use it

1. Add your images. Drop them on the panel above, pick files from your device, or paste from the clipboard. Several at once is fine.
2. Set the target size in KB. Use the exact number your form asks for. If the form gives a range, target the top of it.
3. Check the result. You get the final file size and the dimensions, so you can confirm both against the form rules before uploading.
4. Download. The output is a standard .jpg that government and exam portals accept.
5. Keep your original. Compression is one way, so archive the full-quality file in case the form is rejected and you need to try different dimensions.

Why it matters that this runs in your browser

Nearly every image compressor you will find for this job uploads your file to a server, compresses it there, and sends it back. That means your passport photo, your signature, and your identity documents sit on somebody else's disk, governed by a privacy policy you did not read, usually with a retention window measured in hours or days.

This tool never does that. Your browser decodes the image, re-encodes it through the canvas API, and hands you the bytes back. No request carries the image anywhere. Three practical consequences follow:

• Privacy. Identity documents and photographs never leave your device, so there is no copy to leak, log, or subpoena.
• No limits. There is no bandwidth or CPU bill behind it, so there is no reason to cap you at three files a day or stamp a watermark on the output.
• Speed. A photo compresses in well under a second because it never makes a round trip, which matters when you are working through a multi-document application on a slow connection.

Re-encoding also strips EXIF metadata, so GPS coordinates and camera details do not travel with the file you upload.

Exam forms, government portals, and job applications

Hard KB caps exist mostly because portals were built for slow connections and small storage, and the rules stuck. If you are applying to SSC, UPSC, a state public service commission, an IBPS bank exam, a university admission portal, a PAN or passport service, or a visa application, you have probably met one. The pattern is consistent:

• Photograph. Commonly 20 KB to 50 KB, usually with a fixed pixel size such as 200 by 230.
• Signature. Commonly 10 KB to 20 KB, often 140 by 60 pixels, scanned in dark ink on white paper.
• Documents. Marksheets, certificates, and identity proofs typically get 100 KB to 500 KB each.
• Minimums. Many forms set a floor as well as a ceiling. A photo that is too small is rejected just as fast as one that is too large, so aim near the top of the allowed range.

Always read both numbers on the form, the KB cap and the pixel dimensions, because they are checked separately and a file can pass one and fail the other.

What each target size can actually hold

• 20 KB. A passport-style headshot at 200 by 230 pixels, or a portrait crop up to about 400 by 400. No readable text.
• 50 KB. A 600 to 800 pixel portrait that looks essentially untouched, or a legible photo of a document at 1024 by 768.
• 100 KB. A 1000 to 1200 pixel photo with no visible loss, or an A4 page scanned at 150 dpi in grayscale.
• 200 KB. A 1600 by 1200 photo at near-original quality, or an A4 page at 200 dpi with small print still readable.
• 500 KB. A 4 megapixel photo that is indistinguishable from the original. Usually an email or web performance limit rather than a form rule.

Why JPEG, and when not to use it

For upload forms the answer is almost always JPEG with a .jpg extension. Two reasons. The first is acceptance: most portals allow only JPG or JPEG, some also allow PNG, and very few accept WebP, AVIF, or the HEIC format that iPhones save by default. The second is control: JPEG is lossy, so its quality can be dialled in fine increments until the file lands on a specific byte count. PNG is lossless, so the only ways to shrink it are fewer pixels or fewer colours, which makes an exact KB target awkward.

PNG is still the better choice in one case: screenshots, logos, and line drawings with large flat areas and sharp edges. JPEG smears those edges. If the destination accepts PNG and the image is not a photograph, prefer PNG. WebP is 25 to 35 percent smaller than JPEG at matching quality and is a good choice for your own website, but treat it as unsupported on any government form until you have seen it accepted.

When the target cannot be met cleanly

Sometimes the file fits the cap but looks bad. That is a signal that the image is carrying more pixels than the byte budget can pay for. Work down this list in order:

1. Crop first. A photo that is 70 percent background wall spends 70 percent of its bytes on the wall. Cropping to head and shoulders is the cheapest quality win available.
2. Reduce the dimensions. Halving the width and height quarters the pixel count and gives every remaining pixel four times the budget. A sharp small image beats a blocky large one at the same file size, every time.
3. Go grayscale for scans and signatures. Dropping the colour channels can halve the size of a document image with no loss that matters.
4. Rescan at a lower dpi. A 300 dpi colour A4 scan is roughly 8 megapixels and will not fit a tight cap in any usable state. Rescan at 150 to 200 dpi instead of compressing it into mush.
5. Protect resolution on text. For scanned pages, let quality fall further before you let the dimensions shrink. Letters roughly 20 pixels tall stay readable through mild artefacts; letters 6 pixels tall are gone whatever the quality setting.
6. Reshoot against a plain background. Busy, noisy, or low-light backgrounds are expensive to encode and are against most form rules anyway.

Frequently asked questions

Is my image uploaded to a server?

No. The image is decoded and re-encoded by your own browser using the HTML canvas API. It never leaves your device, nothing is uploaded, and there is no copy for us to store or delete. That is the main difference between this tool and almost every other image compressor in search results.

How does it hit an exact file size in KB?

It re-encodes the image as JPEG at a given quality, measures the resulting bytes, and repeats with a binary search until the file lands just under your target. If quality alone cannot get there, it steps the pixel dimensions down and searches again. You get the largest, sharpest file that still fits the limit.

Will the result be exactly the size I asked for?

It will be at or just under your target, never over. Hitting a byte-exact size is not possible with JPEG because the encoder decides where the bits go, so the tool aims for the closest fit below the cap. That is what upload forms actually check.

Does KB mean 1024 bytes or 1000 bytes here?

1 KB is treated as 1024 bytes, which is what Windows, most upload validators, and government portals report. If a portal happens to use 1000 bytes, a file that fits the stricter 1024-byte reading also fits the looser one, so you are safe either way.

Which format should I choose for a government or exam form?

JPEG, saved with a .jpg extension. Most portals accept only JPG or JPEG, some also accept PNG, and very few accept WebP, AVIF, or HEIC. JPEG is also the only common format whose quality can be dialled down smoothly enough to hit a specific KB target.

My photo came from an iPhone and the portal rejects it. Why?

iPhones save photos as HEIC by default, and almost no upload form accepts HEIC. Opened in Safari, this tool converts a HEIC file to JPEG and hits your size target in the same pass. Chrome and Firefox ship no HEIC decoder at all, so on Windows or Android the file will not open here: switch the iPhone camera to Most Compatible in Settings, Camera, Formats and retake the photo, or convert it on the phone first.

Can I compress several images at once?

Yes. Add multiple files and each one is compressed to the same target independently, then downloaded. This is the fast path when a form wants a photo, a signature, and three certificates all under the same cap.

Is there a file size, resolution, or usage limit?

No usage limit, no signup, no watermark, and no daily cap: run as many images through it as you like. There are two practical caps on a single batch, both there to protect your device rather than a server: 30 MB per file and 20 files at a time. Anything larger is simply added in a second pass.

Does it work on a phone?

Yes. It runs in mobile Safari and Chrome the same way it runs on desktop, which is useful because most exam-form uploads are done from a phone. Large scans compress a little slower on older phones but still work.

Will compressing an image reduce its dimensions?

Only if it has to. The tool first tries to reach your target by lowering JPEG quality at the original pixel size. It reduces the dimensions only when quality alone cannot get under the cap, because a smaller sharp image looks far better than a full-size image crushed to the same number of bytes.

Does compressing remove EXIF data like location?

Yes. Re-encoding through canvas drops EXIF metadata, so GPS coordinates, camera model, and timestamps do not travel with the file you upload. That is a useful side effect when you are sending a photo to a portal or a stranger.

Can I get my original quality back afterwards?

No. JPEG compression is lossy, so the discarded detail is gone for good. Always keep the original file and upload the compressed copy, especially for documents you may need to re-scan at a different size later.

Why not just take a screenshot to make the file smaller?

A screenshot re-encodes at your screen resolution, which often makes the file larger, not smaller, and it usually adds a PNG wrapper that portals reject. Compressing properly gives you control over both the pixel size and the byte size.