Our review procedure begins with small, known sources rather than an important original: images with recorded dimensions plus visible transparency, hard edges, small text, and gradients, or PDFs with numbered pages and known page counts. We keep the source unchanged for a side-by-side comparison.
For a local-processing check, we open the browser network log before file selection. Selecting, previewing, converting, and downloading should not create a request containing the source file. We consider normal requests for the page, hosting, and specialised PDF libraries separately and disclose them in the privacy notice.
After export, we record the actual file type, pixel dimensions, and byte size. We view transparency over light and dark colours; inspect lossy output at 100% and at its intended display size; check SVG traces for silhouettes, gaps, and small details; and compare PDF output with selected pages, total page count, and file boundaries. When the destination is available, we also open the result in the software that will receive it.
We treat this as a functional review, not a universal codec benchmark or a promise that every browser, device, source variant, or third-party platform will behave identically.