vs.
Side-by-side comparisons. Same task, different tradeoffs — usually on privacy, signup / rate limit, and free-tier ceiling. Not a hit piece on any competitor; some do what they do very well. But if privacy and no-signup matter to you, these compare the two on paper.
Choose a comparison
@nap.codesvsTinyPNG
Image compression — MozJPEG/WebP/AVIF in the browser vs TinyPNG's Panda API.
@nap.codesvsremove.bg
AI background removal — ONNX in browser vs remove.bg's server API.
@nap.codesvsiLovePDF
PDF toolkit — merge, split, rotate, fill, redact — in-browser vs uploaded.
@nap.codesvsQRCode Monkey
Custom QR — module shapes, colors, gradients, logo overlay, decoder.
@nap.codesvsSVGOMG
SVG optimizer — SVGO in your browser, no cross-origin config.
@nap.codesvsJSONLint
JSON formatter + validator — native parser, no server hop.
@nap.codesvsSquoosh
Image compression — bulk mode + folder structure vs Squoosh's one-at-a-time.
@nap.codesvsFavicon.io
Favicon bundle — every PWA/iOS/Android/tile size + manifest snippet.
The shared tradeoff
Every online tool that runs on someone else's server has to make three tradeoffs: they pay for the CPU (so free tiers get capped), they see your files (so privacy depends on their word), and they add a network round-trip (so batch jobs are slow). Client-side tools flip all three: your CPU runs the work, your files don't move, and the "server" bottleneck disappears. In exchange, huge files can strain your device, and some tools (LLM-scale AI) can't fit in a browser at all — /cutout/, /watermark-remover/, and /tagger/ are the current heavyweights and each takes a one-time model download.
What we're not claiming
These pages don't argue @nap.codes wins on every axis. TinyPNG has a decade of compression research and a paid enterprise plan. remove.bg models are trained on far more data than any browser-loaded ONNX. iLovePDF has OCR and e-sign flows we don't ship. What we do say: if the file is under a few MB, you don't want to sign up, and you'd rather nothing be uploaded, the browser-only path is usually enough.