How CompNivi Works
Built to get you from "a file that's too big" to "a file that fits" in as few steps as possible, without an account and without sacrificing more quality than necessary.
Choose a tool
Pick a preset target size (like 100 KB) or a custom-size tool if your limit doesn't match a preset.
Upload your file
Drag and drop, or tap to browse from your device. Nothing uploads until you choose a file.
Set your target
Confirm or adjust the target size in KB or MB, or pick a compression level for PDFs.
Preview and download
Check the before/after comparison and file size, then download — or process another file.
The image compression algorithm
For target-size tools, the compressor first tries adjusting JPEG or WebP quality using a binary search: it repeatedly tests quality settings to find the highest one whose output size is at or below your target. Only if quality reduction alone can't reach the target does it reduce the image's pixel dimensions, then repeats the quality search at the smaller size — quality first, dimensions second, to stay as close to the original resolution as possible. Our quality-preserving Compress Image tool skips this search entirely and always keeps full dimensions with one high-quality encoding pass.
The PDF compression approach
For PDFs, size is driven almost entirely by embedded images. The compressor extracts each embedded image, recompresses it using the same quality-search approach (bounded by whichever compression level you choose), and rebuilds the PDF with a more compact internal structure. Unnecessary metadata is stripped along the way. Text, fonts, vector graphics and page layout are never rasterized or altered.
The image upscaler
The 4K Image Upscaler enlarges images in repeated 2x steps using high-quality resampling, then applies a sharpening pass to counteract the softness scaling introduces. It is a classical algorithm, not a generative AI model — it enhances existing detail rather than inventing new detail.
Privacy by default
Image tools and the upscaler run entirely in your browser and never touch our servers. PDF compression is the one exception — it requires temporary server-side processing, and files are held only in memory and discarded immediately after your compressed file is returned.