Compress JPEG
Shrink JPG and JPEG photos with a quality slider or an exact target size in kilobytes.
Compress JPEGThe file containing your coordinates never leaves this device — which is rather the point.
JPG, JPEG, PNG and WebP files are accepted. Paste from clipboard with Ctrl+V (Cmd+V on a Mac).
Keep the original format, or convert on the way out to compare which one is smallest.
The purpose of this page. Removes camera model, lens, exposure, timestamps and GPS coordinates. Turning it off leaves the data in place.
Clean the metadata without changing the pixel size.
Ready. Add a file to begin.
Quality mode · same as input output · 0 files
No files yet. Results will appear in this table as soon as you add one.
Your files are processed by your own browser. Nothing is uploaded to a server, so nothing can be stored, read or shared.
More tools
Each one runs in the browser the same way, with no upload and no account.
Shrink JPG and JPEG photos with a quality slider or an exact target size in kilobytes.
Compress JPEGReduce PNG screenshots and graphics while keeping transparency and hard edges intact.
Compress PNGRe-encode WebP images at a lower quality, or convert JPEG and PNG into WebP.
Compress WebPChange pixel dimensions by width, height or percentage, with the aspect ratio locked.
Resize an imageCombine several photos into one PDF, reorder the pages and choose paper size and margin.
Build a PDFWrite a formatted document in the browser and export it to PDF, DOCX or HTML.
Open the editorSpecification
Every capability written out as text, including the ones where the answer is no. A specification that only listed the wins would not be worth reading.
| Attribute | This tool |
|---|---|
| Supported input formats | JPG and JPEG |
| Output formats | JPEG, PNG, WebP |
| Compression type | LossyJPEG discards detail permanently by design. PNG output is lossless but far larger for photographs. |
| Quality control | Adjustable, 10 to 100Maps directly to the encoder's quality setting. Moving it genuinely changes the output. |
| Exact output file size | Yes — target in KBThe encoder searches for the highest quality that fits under your limit, then reports the size it actually reached. |
| Resize / image dimensions | Yes — max width and heightAspect ratio preserved. The resizer page adds exact pixel and percentage modes. |
| Batch / bulk processing | UnlimitedFiles are encoded in parallel across a Web Worker pool and can be downloaded as one ZIP. |
| Maximum file size | No limit set by this siteBounded in practice by your device's memory, since each image is held uncompressed while it is processed. |
| EXIF and metadata | Stripped by defaultCamera model, timestamps and GPS coordinates are removed. You can switch it off. |
| Transparency / alpha channel | Not applicable to JPEGJPEG has no alpha channel. Choose PNG or WebP output if you need transparency. |
| Chroma subsampling control | Not availableThe browser's encoder chooses it and exposes no setting. We will not add a control that cannot affect the output. |
| Progressive JPEG control | Not availablecanvas.toBlob gives no control over progressive encoding. Use MozJPEG or ImageMagick if you need it. |
| Colour profile handling | Converted to sRGBA browser canvas works in sRGB, so wide-gamut images are converted. Correct for the web, wrong for print masters. |
| Bit depth | 8 bits per channelThe maximum a browser canvas provides. HDR and 10-bit output need AVIF, which browsers cannot write. |
| Animation support | Not supportedA canvas decodes only the first frame of an animated file, so animation is out of scope rather than half-supported. |
| Where processing happens | In your browserNo upload endpoint exists in the compression path. Disconnect from the network after the page loads and it still works. |
| File retention | None — nothing is receivedThere is no copy to retain or delete, because no file is ever transmitted. |
| Watermark | NeverThe output contains your image and nothing else. |
| Account, quota or cost | NoneNo sign-up, no daily cap, no paid tier withholding a feature. |
| Browser support | Chrome, Edge, Firefox, SafariDesktop and mobile. Requires Canvas and Web Workers, which every current browser provides. |
How it works
Drag them in, click to browse, or paste from the clipboard. Nothing is transmitted.
Stripping metadata means re-encoding, so pick a quality. 92 is high enough to be invisible.
The output carries no camera model, no timestamps and no coordinates.
Features
Queue as many image files as you like. A Web Worker pool handles them in parallel so a large batch does not freeze the page.
The quality slider maps directly to the encoder setting. Move it and the output genuinely changes — there are no cosmetic controls here.
Ask for 200 KB and the encoder searches for the highest quality that fits under your limit, then reports what it actually achieved.
Camera model, lens, timestamps and GPS coordinates can be removed on the way out. Useful before publishing a photo taken on a phone.
Write the result as JPEG, PNG or WebP regardless of what you put in, so you can test which format is smallest for a given image.
The whole pipeline runs in your browser through the Canvas API. Turn off your network connection after the page loads and it still works.
The output is your image and nothing else. There is no account, no daily quota and no paid tier holding back a feature.
The layout, the drop zone and the results table are all usable on a phone, and mobile browsers run the same encoder as the desktop ones.
In depth
A photograph taken on a phone carries a metadata block alongside the pixels. It typically records the camera make and model, the lens, aperture, shutter speed, ISO and whether the flash fired; the date and time to the second; the orientation the camera was held in; the software the file has passed through; and, if location services were enabled, the latitude, longitude and altitude where the shutter was pressed.
Some files also embed a thumbnail preview, which has a long history of still showing the image as it was before it was edited.
GPS coordinates in a photograph are typically accurate to a few metres — enough to identify a specific house. People photograph things at home constantly: an item for sale, a pet, a document, a room. When that file is posted publicly, the coordinates go with it, and anyone who downloads it can read them in seconds with free tools and no expertise.
Large social platforms strip metadata on upload, which has made this feel solved. It is not. Files sent by email or messaging apps, uploaded to forums and smaller sites, attached to marketplace listings or shared in cloud folders very often keep everything.
There are two ways to remove metadata. One is to edit the file's headers in place, leaving the compressed image data untouched — genuinely lossless, but it requires parsing and rewriting each format's container by hand. The other is to decode the image to raw pixels and write a completely new file, which cannot carry anything across because nothing copies it.
This tool does the second, because it is what a browser can do reliably across every format. The trade-off is a re-encode, which is why there is a quality setting here at all. At 92 the change is invisible; the honest thing is to tell you it happened rather than to imply the file was untouched.
One benefit of the re-encode: orientation is baked into the pixels before saving, so a portrait photograph cannot end up sideways the way it can when an orientation tag is deleted carelessly.
Better than cleaning files afterwards is not recording the location in the first place.
Both can be switched on for a trip and off again afterwards. Keep EXIF in your personal library, where the dates make it searchable years later; strip it from everything you publish or send.
Questions
Every answer below is present in the page source, expanded, so it can be read without opening anything.
Yes. The image is decoded to raw pixels and written as an entirely new file, so nothing from the original metadata block is carried across — including latitude, longitude and altitude. There is no partial strip and no field left behind.
It is re-encoded, which for JPEG and WebP means a lossy pass. At the default quality of 92 the difference is invisible, but it is a real change and we would rather say so. If you need the pixels untouched, convert to PNG output, which is lossless.
No. The orientation tag is applied when the image is decoded, so the rotation is baked into the pixels before the new file is written. That is a common failure with tools that simply delete the metadata block and leave the pixels as they were.
Not on this page — it removes metadata rather than displaying it, and we would rather ship one tool that works than a viewer that half-parses a dozen formats. Most operating systems show the main fields in the file properties or info panel.
It would be, which is exactly why nothing is uploaded here. The file is read into this browser tab from your own disk, cleaned in memory and written back to your device. The version containing your coordinates never crosses the network.
Yes, in your own library. Photographers use it to review which settings produced which result, dates make an archive searchable, and stock and press submissions often require it. The rule of thumb is to keep it privately and strip it from anything you publish.
A little — usually a few kilobytes, occasionally more when a large embedded thumbnail or colour profile is present. It is a genuine saving, but it is not the reason to do this. Privacy is.
Related topics
Related concepts, formats and guides across the site. Every link goes to a page that covers the term properly.
From the blog
Longer explanations from the blog.
A photograph from a phone can carry the coordinates of the place it was taken, the camera serial number and the exact timestamp. Removing that data is straightforward. Removing it carelessly rotates your photos sideways.
9 min read
Resize before you compress. Measured on one photograph: half the width at quality 80 produced a 16.4 KB file, while full width at quality 40 produced 20.6 KB and looked far worse. The order of operations matters more than the settings.
8 min read
Attachments are base64-encoded in transit, which inflates them by about a third. A 20 MB limit is really a 15 MB limit. Resize to 1600 pixels on the long edge and encode at quality 80, and a photograph lands near 150 KB.
9 min read