To compress image files effectively, begin with the limit or destination you need to satisfy. A form that requires a file under 500 KB is a different task from a website that needs several photographs to load faster. This page lets you process a batch of JPG, PNG and WebP images locally, adjust quality and dimensions, then inspect the actual output before downloading it.
The useful result is not simply “smaller.” It is a file that is small enough for the destination while its important details still look acceptable. Faces, product edges, text, gradients and transparent areas deserve different checks, so this tool shows both the preview and measurable file information instead of hiding the trade-off behind one automatic score.
Start with the upload requirement
Read the destination instructions before changing a file. Requirements usually describe one or more of four things: file size, pixel dimensions, format or transparency. They are connected, but they are not interchangeable.
| Requirement | Control to use on this page | What to verify |
|---|---|---|
| Maximum KB or MB | Quality, format and optional size limit | Actual compressed size |
| Maximum width or height | Resize large images | Output dimensions |
| JPG, PNG or WebP required | Output format | Reported MIME type and filename |
| Transparent background needed | Keep PNG or choose WebP | Transparency in the compressed preview |
| Many files need the same rules | Add a batch and use one group of settings | Each item in the result queue |
This compressor reports the output it actually created. It does not promise an exact percentage reduction because two files with the same dimensions can compress very differently. A smooth photograph, a noisy night scene and a screenshot full of small text contain different visual information.
How to compress images on this page
- Add your images. Select or drop as many as 30 JPG, PNG or WebP files. The page checks format, file size and browser-safe dimensions before adding them to the queue.
- Choose an output format. Keep each original format, or convert the entire batch to JPG, WebP or PNG. Remember that JPG cannot retain transparency.
- Set quality. JPG and WebP use the quality percentage as an encoding request. Start near 80–85% for general photographs, then inspect detailed areas. PNG output remains lossless and does not use this slider.
- Limit dimensions when useful. If an image is much larger than its final display, choose a maximum longest side. Smaller pixel dimensions can save more data than repeatedly lowering quality.
- Compress and inspect. The result panel compares original and output previews, total batch size, per-file savings, dimensions, format and transparency behavior.
- Download the result. Download the selected image, or package multiple successful results into one ZIP file.
The knowledge in this guide maps directly to the controls above: use Quality for the encoding trade-off, Output format for compatibility and transparency, and Resize large images when excess pixel dimensions are the main source of file weight.
Choose quality, format and dimensions
Quality is a trade-off, not a universal score
Lower JPG or WebP quality usually reduces file size by discarding more visual information. The percentage is an encoder input rather than an objective measure of how good the output looks. Browser implementations and image content influence the final result, so compare the preview at the size where the image will really be used.
Look closely at eyes and hair in portraits, product outlines, lettering, repeated textures and smooth color transitions. If those areas look smeared, blocky or surrounded by halos, raise quality. If the output still looks clean and the file is too large, lower quality gradually rather than making one extreme change.
JPG, PNG and WebP serve different jobs
- JPG usually suits photographs when broad compatibility matters and transparency is unnecessary. This page fills transparent pixels with white before creating JPG output.
- PNG suits transparency, interface graphics, diagrams and sharp text edges. Re-encoding a detailed photograph as PNG can create a larger file because this page uses lossless PNG output.
- WebP is useful for modern web delivery when the destination accepts it. It supports the quality control on this page and can retain transparency.
- Keep each format applies the same batch settings while retaining each source type. JPG and WebP use the selected quality; PNG remains lossless.
Dimensions often matter more than another quality reduction
A 6000-pixel photograph displayed at 1200 pixels wide contains far more pixel data than the page can show. Limiting its longest side can substantially reduce the output while keeping a reasonable quality setting. The size presets on this page only shrink images that exceed the selected limit; they do not enlarge smaller files.
Compression and resizing solve different problems. When a destination mixes a file-size limit with exact width or height requirements, satisfy the dimensions first and then inspect the compressed file size.
Read the batch result
Start with the overall totals, then inspect individual files. The batch summary shows original bytes, compressed bytes and the net saving. A negative saving means the output is larger; it is a signal to change settings, not a processing error.
For the selected result, check:
- Preview: compare important visual details in the original and compressed versions.
- File size and savings: use the generated byte count rather than an estimate.
- Dimensions: confirm that an optional maximum-size rule did not make the image too small for its destination.
- Format and MIME type: make sure the receiving website accepts the output.
- Transparency: JPG is flattened on white; PNG and WebP retain transparency when the browser encoder supports it.
- Metadata: files generated by this page do not copy the source EXIF or other embedded metadata.
The batch can contain mixed outcomes. One photo may shrink substantially while a small, already optimized logo becomes slightly larger. Select each result in the queue when the destination is strict or the image contains text and fine detail.
Illustrative batch example
Suppose you add three large JPG photographs totaling 18 MB. You keep JPG output, choose 82% quality and limit the longest side to 1920 pixels. The resulting batch might look like this:
| File | Original | Illustrative output | Main reason for reduction |
|---|---|---|---|
| portrait.jpg | 6.4 MB | 1.1 MB | Fewer pixels and re-encoding |
| storefront.jpg | 7.2 MB | 1.5 MB | Fewer pixels and re-encoding |
| product.jpg | 4.4 MB | 980 KB | Fewer pixels and re-encoding |
Illustrative example only. Actual file sizes depend on source pixels, image content, existing compression, output format, browser encoder and selected quality. Use the result panel for the real values produced on your device.
Where base97 differs from other compressors
This page is designed around an explainable batch workflow rather than a single automatic optimization button. It processes up to 30 images locally, exposes format, quality and maximum-dimension controls, shows original and compressed previews, reports per-file results and creates a ZIP in the browser.
That does not make it the right tool for every case:
- TinyPNG provides a highly automated upload workflow and supports additional input such as AVIF. Its public free tool states limits of 20 images and 5 MB per image. base97 accepts fewer formats, but its current batch limit is 30 images with up to 40 MB per file and the files stay in this browser.
- iLoveIMG supports JPG, PNG, SVG and GIF plus cloud imports. Choose it when those formats or Google Drive and Dropbox are important. Choose base97 when local JPG, PNG or WebP processing and visible encoding controls matter more.
- Adobe Express combines compression with a broader editing workflow and accepts supported images up to 40 MB. base97 is more focused: one batch, explicit settings, result inspection and direct downloads without moving into a design editor.
- Squoosh is a strong option for detailed, per-image codec experimentation. base97 provides fewer advanced codec parameters but is organized for applying consistent settings across a multi-file queue.
The browser creates output through the Canvas API. As documented for toBlob(), format and quality support depend on the browser. This page verifies the returned MIME type and reports an error rather than silently labeling an unsupported format as successful.
Troubleshoot image compression
The result is larger than the original
The source may already be optimized, the quality may be high, or the selected format may be inefficient for that content. Lower JPG or WebP quality, choose WebP when the destination accepts it, or limit oversized dimensions. Avoid converting a photograph to lossless PNG when a small file is the main goal.
Text or edges look blurry
Raise quality, keep PNG for interface graphics, or avoid reducing dimensions below the final display size. Small text and thin lines reveal compression damage earlier than many photographs.
Transparent areas became white
JPG has no transparency in this workflow. Choose PNG or WebP output and inspect the compressed preview before downloading.
A file was skipped
This page accepts JPG, PNG and WebP files up to 40 MB each, with a 200 MB batch total. It also rejects images beyond 12,000 pixels on either side or 60 megapixels to reduce browser memory failures. HEIC, AVIF, SVG and animated GIF files are not supported here.
The browser runs out of memory
Large decoded images use much more memory than their compressed file size suggests. Try fewer images in one batch, choose a maximum dimension, close memory-heavy tabs or process the largest file separately.
Common questions about compressing images
Does this image compressor upload my files?
No. The selected JPG, PNG and WebP files are decoded, redrawn and encoded by browser APIs on your device. This tool does not send them to a base97 server.
Can I compress an image to an exact size such as 100 KB?
Not on this page. You can adjust quality, output format and maximum dimensions, then check the actual result. The compressor does not repeatedly target an exact KB value.
Why did a compressed image become larger?
A high quality setting, a less efficient output format or an already optimized source can produce a larger result. Compare the reported sizes, lower quality, choose WebP for a compatible destination or reduce large dimensions.
Does compression remove photo metadata?
Yes for files produced by this page. The browser draws image pixels to a canvas and creates a new file, so original EXIF and other embedded metadata are not copied into the output.
Research and editorial notes: The base97 implementation and limits were reviewed on September 27, 2026. Browser encoding behavior was checked against MDN Canvas documentation. Competitor scope was reviewed from the official TinyPNG, iLoveIMG, Adobe Express and Squoosh pages. Comparisons describe workflow differences and do not claim that one encoder always produces a smaller or better-looking file.