About UtilTools

Tools for the things you actually need to do, built on one rule: if it needs a server, it does not go in.

The observation this started from

Most online tools upload your file. Not because they have to — because it was the easy way to build them, and because a file on a server is an asset. iLovePDF handles something in the order of 217 million visits a month. Smallpdf around 40 million. Every contract, every scanned passport, every payroll spreadsheet run through those sites is copied onto a machine belonging to someone else, governed by a retention policy nobody reads.

For a long time that was simply how it worked. Rendering a PDF page, re-encoding a JPEG, deriving an encryption key — those were server jobs. They are not any more. A browser in 2026 has Canvas, Web Crypto, WebAssembly, Compression Streams, the File System Access API and Web Workers. The work can happen on the device that already has the file.

The rule

A tool goes in the catalog only if it runs entirely on your device. That single constraint decides everything else here: what gets built, what gets refused, what the infrastructure costs, and what the site can honestly claim.

It is a real constraint, not a marketing line. It is why there is no PDF to Word converter, no translator, no currency converter and no speed test — all of them have more demand than most of what is here. We publish the full list of what we refuse to build and why, because a promise with a single exception is not a promise.

What that means in practice

• No account. There is nothing to attach an account to.
• No daily limit. Limits exist to ration server capacity. There is no server capacity.
• No watermark. Watermarks exist to sell the paid tier. There is no paid tier.
• No file size tier. The limit is your device's memory, and each tool tells you what it is.
• It works offline. Install it as an app and the tools keep working with the network off — which is the clearest demonstration that nothing was being uploaded.

How it is built

Every page here is generated at build time. One folder per tool, each with its complete text content in the served markup, because AI crawlers do not execute JavaScript and a page that renders empty without JS does not exist as far as they are concerned. Plain CSS with custom properties — no framework, no runtime compilation. No third-party scripts, no CDN, no analytics that follows you.

The tool logic runs entirely in your browser with no external dependencies. The PDF engine, the QR encoder, the EXIF parser, the ZIP writer, the diff algorithm, the expression parser — all of it is written for this project. That is more work than pulling in libraries, and it buys three things: a strict Content Security Policy, genuine offline capability, and no supply-chain surface at all.

Who builds it

One developer: Gabriel Masson. That matters more than any privacy badge: the claim on every page is that your file never leaves your device, and you can open your browser's network inspector and watch nothing happen.

How it is paid for

A static site costs almost nothing to host, which is what makes this model possible. What cost there is, and the time behind it, is covered by clearly disclosed partner recommendations on the Partners page. No tool is influenced by a partnership, no tool is gated, and no partner has any say over the catalog. If that ever stops being true, it will say so here first.

What is coming

Two tools are missing that people ask for constantly and that can run in a browser, just expensively: optical character recognition and background removal. Both need a model download in the tens of megabytes. They will ship when they can be offered with the cost stated up front rather than sprung on someone on a mobile connection. Video conversion is a harder case, and the reasoning is on the excluded list.

Questions

Last updated 2026-08-22.