О Absolutool

Что такое Absolutool?

Коллекция из более чем 240 бесплатных онлайн-инструментов, полностью работающих в вашем браузере. Обрабатываемые вами файлы никогда не покидают ваше устройство · без загрузок, без регистрации, без учётной записи.

Принципы

Реклама и аналитика

Absolutool использует Google AdSense для показа ненавязчивой рекламы (без всплывающих окон, промежуточных экранов и платного доступа) и Google Analytics для понимания, какие инструменты используются. Ни один из этих сервисов не видит обрабатываемые вами файлы · они остаются в вашем браузере. Полное раскрытие и ссылки на отказ на странице Политики конфиденциальности.

Контакты

Есть предложение, обнаружили ошибку или идея для инструмента? Напишите нам на absolutoolweb@gmail.com. For more on what we can help with and typical response times, see the Contact page.

Why we built this

Most "free online tool" sites work the same way: you upload a file, it gets processed on a remote server, you download the result, and the site quietly keeps a copy of your data on disk for some indeterminate period. Even sites with the right intentions concede they have to upload your file somewhere — to a Lambda, a worker queue, a temporary S3 bucket — and that moment of upload is when the privacy promise becomes a matter of trust rather than a matter of architecture.

Absolutool exists because it doesn't have to work that way. Modern browsers can do almost everything those server pipelines do, locally, on your device, with no upload at all. Image resizing, PDF splitting, video trimming, base64 encoding, password generation, regex testing, color contrast checking, EXIF inspection, OKLCH gradient previews — every one of these can run inside the browser tab with no network round-trip. The site started as one person's frustration that no general-purpose tool site was actually being honest about this and shipping the client-side version. The 248 tools currently live are the result.

Editorial standards

Every tool page on this site has a body-copy section explaining what the tool does, the history and standards behind the format, and the specific decisions that informed its implementation. Those sections aren't auto-generated. Each one was written by hand against a per-tool research dossier saved to the repository before the copy was written, so the published facts have a citable trail. The dossiers cover things like the RFC that defines a format, the academic paper that established a result, the historical timeline of a specification, and the failure modes that a careful implementation has to handle. When research surfaced a contradicted or unverified claim, the copy either omits the claim or flags it openly.

This matters because the easiest way to fill a tool site with words is to ask a language model to generate three paragraphs of plausible-sounding filler per page. That kind of content is detectable, dilutive of the rest of the web, and mostly worthless to an actual reader. The opposite — slow, methodical, source-traced writing — is the only way the site can earn the modest amount of attention it asks for.

How translations work

The site ships in twenty languages: English, French, Chinese, Spanish, Hindi, Portuguese, German, Italian, Japanese, Indonesian, Vietnamese, Korean, Arabic, Russian, Turkish, Polish, Dutch, Thai, Swedish and Ukrainian. Translations are added by hand, string by string, by the site's owner. There are no auto-translated pages and no LLM-generated localisations on the live site. New languages are added only when their translation files reach parity with the English source — which takes a while per language, but means a French or Italian visitor lands on a page that reads as if it was written in French or Italian, not on machine output with the seams showing.

If a language switch on this page produces something that reads awkwardly, that's a translation bug to report (use the email address above). The standard the site holds itself to is that any translation good enough to ship should be good enough to read out loud without wincing.

How "client-side only" actually works

When you open a tool here, your browser downloads HTML, CSS and a small amount of JavaScript — the same way it downloads any web page. The JavaScript then does the work locally, using browser APIs that have existed for years: FileReader for reading files into memory, the Canvas API for image manipulation, window.crypto.getRandomValues for random number generation, WebAssembly for compiled libraries like FFmpeg and pdf.js, and IndexedDB for the optional Saved Files vault. None of these APIs send anything to the server.

You can verify this yourself by opening your browser's network tab (F12 in most browsers, then Network) before processing a file. Other than the initial page load, there are no upload requests. The server doesn't have a route to accept your file even if it wanted to — there's no upload endpoint at absolutool.com. The same architectural constraint that makes the site fast also makes the privacy promise straightforward to audit.

Who maintains the site

Absolutool is run by an independent solo maintainer. There is no investor, no parent company and no team of editors signing off on each release. The site is hosted on Cloudflare Pages, the source repository is public on GitHub, and the build runs on every commit. Reader-funded only in the sense that ad revenue (when AdSense approves the site) covers hosting and a small amount of the maintainer's time.

The decision to build a 240+-tool site as a one-person project rather than a venture-backed product is intentional. A small site can credibly promise no tracking beyond what's listed on the privacy page, no upsell tier, no email-capture interstitial, and no deprecation of "premium" features behind a paywall introduced after launch. Those promises are easier to keep when there's no growth target to satisfy.

Reporting issues and requesting tools

The fastest way to report a bug is the floating feedback button on the bottom-right of every page (the speech-bubble icon). The fastest way to request a new tool is the same button, set to "Suggestion / Idea." Both deliver to the same inbox the contact email points to. There's no public roadmap because the answer to "what gets built next" is mostly "what readers actually ask for, weighted by how quickly it can be built without compromising the no-server constraint."

The repository on GitHub is open for issue reports if you'd rather use that — link in the Contact page — but the email is checked daily and is the right channel for non-developers.