Why editing in the browser is safer than uploading
Most online PDF and image tools work the same way. You choose a file, it travels over the internet to a server you have never heard of, a program you cannot inspect changes it, and a new file travels back. The interface makes this feel local. It is not local. For the minutes or hours that file sits on that machine, its contents are readable by the company that runs it, by anyone who compromises it, and by anyone who can compel the company to hand it over.
For a holiday photo, none of that matters. But look at what people actually put through these tools. Signed tenancy agreements. Passport and driving licence scans for a visa application. Payslips for a mortgage broker. Medical letters. Divorce paperwork. Board minutes. Bank statements with an account number on every page. The average "merge two PDFs" job is far more sensitive than the average email, and it gets treated with far less care.
What "client-side" actually means here
Modern browsers can do the work themselves. The File API reads the bytes of a file you have chosen without sending them anywhere. Canvas and WebAssembly can decode, transform and re-encode images and PDFs at speeds that were server-only work a decade ago. Web Workers keep that computation off the main thread so the page stays responsive. When the job finishes, the result is handed back to you as a download that was assembled in memory on your own machine.
The practical consequence is that you can open your browser's network panel, run any tool on this site, and watch nothing leave. That is not a promise you have to take on trust — it is something you can verify in about fifteen seconds, which is the only kind of privacy claim worth making. It also means the tools keep working when the connection does not, and that there is no queue, no page cap and no "upgrade to process files larger than 10 MB".
Where the honest limits are
Doing everything locally has a real ceiling and it would be dishonest to pretend otherwise. Very large files are bounded by your device's memory rather than by a rented server's; a 400-page scanned document will process faster on a laptop than on a five-year-old phone. Editing existing PDF text is genuinely hard, because a PDF stores positioned glyphs rather than paragraphs — so the editor here offers three clearly separated modes and tells you which one suits your job instead of silently mangling the layout. Optical character recognition runs on your device, which keeps the scan private but takes longer than a data centre would. Where a limit exists, the tool says so before you start.
Built five times, on purpose
The second half of the site is calculators, and they are localised rather than translated. A take-home pay calculator is not a currency swap away from a paycheck calculator: the United Kingdom has PAYE, National Insurance and several student loan plans; Canada has CPP and EI on top of federal and provincial tax; Australia has PAYG, the Medicare levy, HECS-HELP and the superannuation guarantee; the United States has federal brackets plus fifty different state regimes; and India has two entire parallel tax regimes you choose between, a cess charged on the tax rather than on income, and EPF and state professional tax in place of any single payroll levy. Each of those gets its own page, its own worked example and its own vocabulary, with the rate source and tax year printed on it so you can check whether it is current before you rely on it.
The same care applies to the small things. Dates read MM/DD/YYYY in the US and DD/MM/YYYY in the UK, Australia and India. Fuel economy is miles per US gallon, miles per imperial gallon, litres per hundred kilometres or kilometres per litre depending on where you are. Large numbers group in lakh and crore for an Indian visitor and in thousands for everyone else. PDF exports default to Letter or A4 without asking. None of it is clever — it is just the work that most multi-country tool sites skip.