Home / Images & PDF

Image Compressor

100% client-side processing: your files never leave your browser.

or drop images here. They stay on your device.
70

Choose or drop images to start.

    An image compressor reduces file size by re-encoding images and, when requested, limiting their dimensions. The default Quality setting of 70 offers a useful comparison point for WebP and JPEG, but it is not a universal target. Process a representative image, inspect the result at its actual display size, and keep the smallest version that still meets the visual requirement.

    How to Use the Image Compressor

    Compare compression settings against the original rather than choosing a quality number in isolation.

    1. Select Choose images, or drag and drop multiple files onto the drop area.
    2. Leave Format at Same as original, or select WebP, JPEG, or PNG for a different output.
    3. Set Quality from 10 to 100 for WebP or JPEG. The default is 70; PNG output remains lossless and does not use this setting.
    4. Optionally enter Max width and Max height in pixels. Leave them empty to retain original dimensions; images are only scaled down, never up, and always keep their aspect ratio.
    5. Select Compress to process again with the current settings. Initial processing happens automatically after file selection.
    6. Compare the output file name, dimensions, original size, new size, and percentage change in each result row.
    7. Select Download for one image or Download all for the batch. The browser may ask to allow multiple downloads, and no ZIP is created.

    Lossy and Lossless Image Compression

    Lossless compression preserves the decoded pixel values needed to reconstruct the image. PNG uses lossless compression, which suits interface graphics, diagrams, and images that require exact edges or transparency. Re-encoding a PNG as PNG may change its byte size, but the Quality slider cannot make it progressively smaller.

    Lossy compression discards image information that an encoder considers less visible. JPEG and lossy WebP use this approach. It works especially well for photographs, gradients, and textured scenes where small local changes are difficult to notice. At aggressive settings, smooth areas may show banding, edges may develop halos, and fine texture can look smeared.

    Content First format to test Main risk to inspect
    Photograph WebP or JPEG Texture loss and block artifacts
    UI screenshot PNG or WebP Soft text and edge ringing
    Transparent graphic WebP or PNG Alpha-edge quality
    Flat logo PNG Unnecessary palette or color complexity

    What the Quality Slider Actually Controls

    The Quality value is an encoder input, not a percentage of pixels retained and not a promised file-size reduction. Quality 70 does not mean that 70 percent of the original information remains. Its visible effect varies with format, browser encoder, dimensions, noise, sharp edges, and the source’s prior compression.

    Use a short comparison loop. Begin near the default, inspect the result at the size users will see, then lower Quality until defects become distracting. Move back one step and verify difficult regions such as faces, hair, foliage, gradients, text, and high-contrast boundaries. A large monitor zoomed to 400 percent can reveal differences that never matter in the layout, while a quick thumbnail check can hide defects visible in a hero image.

    Why PNG Files May Stay Large

    PNG compresses repeated pixel patterns without throwing away decoded image data. This is effective for flat graphics with limited colors, but photographs contain many small changes and do not usually compress as compactly. Moving the Quality slider has no effect on PNG because the output remains lossless.

    When a large PNG is actually a photograph, test WebP. WebP can preserve alpha transparency while using lossy color compression, so it also helps with some photographic cutouts. If pixel-exact reproduction is required, stay with PNG and focus on removing unnecessary dimensions rather than expecting a lossy quality control.

    If PNG-to-PNG output grows, retain the original. It may already use palette or filtering decisions the browser encoder does not reproduce.

    Image Weight and Largest Contentful Paint

    Image bytes can affect Largest Contentful Paint when the LCP element is a hero image or other large visual. Fewer transferred bytes can shorten download time, particularly on slow connections, but compression is only one part of the loading path. Discovery timing, server latency, responsive source selection, caching, and decode work also matter.

    Serve an image close to its rendered dimensions. Compressing a 5000-pixel photograph while displaying it at 800 pixels still makes the browser download and decode unnecessary data. Set Max width and Max height when the source is larger than the largest intended presentation, then consider responsive srcset variants for different layouts.

    Privacy, Metadata, Browser Limits, and Review Workflow

    The tool decodes with createImageBitmap, applies EXIF orientation, draws to a canvas, and encodes with canvas.toBlob. This removes EXIF, GPS location, and camera metadata. Canvas output is converted to sRGB, so a wide-gamut or custom-profile source can shift in color.

    Animated GIF, WebP, and APNG files become a single still frame. Transparency is retained by PNG and WebP but becomes white in JPEG output. HEIC and HEIF work only in browsers with native decoding, such as recent Safari versions; Chrome and Firefox currently cannot convert them here. If an older Safari version cannot encode WebP, the tool warns instead of silently producing PNG data.

    Files stay in the browser and are never uploaded. Once loaded, the page also works offline. Local processing consumes memory, so split an unusually large batch when needed.

    Review both visual output and bytes. The result list provides dimensions and percentage change, but it cannot decide whether an artifact is acceptable for a product image, technical diagram, or background texture. That decision belongs to the design and content context.

    Frequently Asked Questions

    What image quality setting should I use?

    There is no universal value. Start around the default Quality of 70 for WebP or JPEG, inspect representative detail at the final display size, and adjust until you find the smallest acceptable output.

    Why does the Quality slider not shrink my PNG?

    PNG output is lossless, so the slider does not apply. Try WebP when lossy compression and transparency are acceptable, or reduce pixel dimensions.

    Can compression make an image larger?

    Yes. An already optimized source or a PNG re-encoded through canvas may produce larger output. The result reports the increase so you can retain the original.

    Are my images uploaded for compression?

    No. Processing happens locally in the browser, files are never uploaded, and the page continues to work offline after it loads.

    Does compression remove EXIF and GPS data?

    Yes. Re-encoding through canvas removes metadata including EXIF, camera information, and GPS coordinates. It also converts color to sRGB and reduces animated inputs to one still frame.