All your everyday little tools, free and private.

PDF, images, text, conversions, SEO, calculators: everything runs directly in your browser. No file uploads, no required sign-up.

40 free tools · Your files never leave your browser.

NexavoTools brings together forty office utilities that run directly in your browser: PDF compression and merging, image conversion and resizing, text recognition, transformation of Word and Excel documents, generators for search engine optimisation, financial calculators and text analysis tools. There is no account to create, no software to install, no watermark on the files produced and no daily limit.

What separates this from an ordinary online converter fits into one sentence: your file is not uploaded. The page loads the engine it needs, your browser reads the file from your own disk, transforms it in memory and offers it back for download. Nothing travels through a server, so there is no queue to wait in, no deletion delay to keep an eye on, and no promise of erasure to take on trust.

In practice that changes the nature of what you can put through these tools. A contract, a payslip, a scanned identity document or a medical record can be processed without leaving your device, which is true of no service that has to receive your documents in order to work on them remotely.

What "in your browser" means in practice

Modern browsers can run compiled code at close to native speed, thanks to WebAssembly. NexavoTools uses that capability to carry real processing engines rather than approximations of them: Ghostscript to recompute a PDF, MozJPEG, OxiPNG and libwebp to re-encode images, Tesseract to read the text inside a photo. All of that code is served by the site itself, from its own assets folder, so what runs is what the site publishes, and your browser asks no third-party network for it. The download happens once, which explains the initial loading time of some tools, and the traffic runs in one direction only: it is the program arriving, not your file leaving.

The trade-off is a real one. The work draws on your device’s memory and processor: a 300 MB PDF compresses noticeably more slowly on a phone than on a desktop machine, and a tab can run out of memory where a server would not have blinked. Files are accepted up to 1 GB, with a warning beyond 400 MB depending on the memory available. Closing the other tabs before a heavy job is usually enough to unblock the situation.

One useful consequence follows: the tools carry on working offline once their engines are loaded, since there is nothing left to ask the network for. The engines themselves are cached by the browser, so the wait happens once per tool rather than once per file.

Three tools are exceptions, and say so

The spell checker sends text to the free LanguageTool service, translation sends it to the MyMemory API, and the SEO audit fetches the address you type in. No browser can do those three things on its own: checking grammar assumes a large grammatical dictionary, translating assumes a model, and reading a page belonging to another domain is forbidden by browser security. The pages concerned state this outright rather than letting you assume the work is local.

The audit has one detail of its own. When the remote site refuses to be read directly, the request goes back out through a public relay, api.allorigins.win, and the result then says the page was read through a proxy. What leaves the browser in these three cases is text, or a web address — never a file. No tool on the site uploads a document.

The same concern for precision applies to the tools filed under "AI". The automatic summary rests on a statistical analysis of sentences, the rephraser applies writing rules, and the title generators work from templates. Only text recognition in an image uses a genuinely trained model, and that one runs locally. Each page describes the method actually employed, which also tells you what to expect: a rule-based tool returns the same result twice for the same input, and invents nothing.

How to choose the right tool

The seven categories above match the type of file you start from, not the result you are after, and that is the point that most often sends people to the wrong page. To bring down the weight of a document for an online form, PDF compression is the right starting point; if the document turns out to be a photograph of a page, converting the image to PDF gives a lighter and cleaner result. To make a scan searchable, what you need is text recognition, not compression.

Every tool page describes the engine used, the formats accepted, what the transformation keeps and what it loses, and the cases in which the tool fails. Several pages also say when you would be better served by something else, a desktop application for instance. A site that sells all of its tools as the right answer to everything is doing nobody a favour.

Free, funded by advertising, open source

NexavoTools is free with no hidden counterpart: no "pro" version, no account, no resale of data, since the site collects none. The funding comes from Google AdSense advertising, and the banner shown on your first visit lets you choose between all ads, a single discreet ad per page, or no advertising at all, in which case no advertising script is loaded. That choice can be changed later from the footer.

The site’s code is published under the AGPL-3.0 licence and can be read by anyone, which makes it possible to verify what each tool really does instead of relying on a statement of intent. The site is published by MATTOIR, a sole trader (entrepreneur individuel) established in Sainte-Marie, on Réunion, a French overseas department; the full identity appears in the legal notice, and the About page sets out the technical choices behind the site.

Where a browser tool stops being the right answer

Some jobs sit outside what a tab does well, and the tool pages say so rather than letting you find out. A password-protected PDF cannot be opened by the pdf-lib tools at all: remove the password in the software that set it before doing anything else. A .docx converted here comes back as page images, so anything meant to be quoted, searched or edited afterwards deserves the native PDF export in Word or LibreOffice. Colour-managed print work belongs in a real imaging application, since the image tools decode to raw pixels and drop EXIF metadata, capture date and GPS coordinates included, on the way out.

Size is the other boundary. Batches of several hundred images, PDFs running to several hundred pages, and audio close to the forty-five-minute cap are the cases where a desktop program finishes while a tab gives up. Splitting the work into smaller runs usually gets a browser through it. Past a certain point the honest answer is to use something else, and each tool page names the point where that happens.

Frequently asked questions

Are my files sent anywhere?

No, for the overwhelming majority of the tools: the processing takes place in your browser’s memory and the file never leaves your device. The three exceptions — the spell checker, translation and the SEO audit — send text or a web address to a third-party service, and each says so on its own page. No tool sends a file.

Do I have to create an account or pay?

No. There is no account, no subscription, no paid tier, no daily quota and no watermark on the files produced. The site is funded by advertising, and you set the level of it yourself in the consent banner on your first visit.

What file size is accepted?

Up to 1 GB, with a warning beyond 400 MB depending on the memory available on the device. The real limit is your browser’s: a recent computer handles without effort what will make an entry-level phone fail. If a job stops halfway, close the other tabs and try again.

Do the tools work on a phone?

Yes. The interface is built for small screens and the local processing works in exactly the same way. The difference is memory: large files and heavy operations, such as text recognition on a document of several pages, need considerably more patience and can fail where a computer would get there.

Why do some tools take a while to start?

They download their engine first, and they are not all the same weight: around 16 MB for Ghostscript behind PDF compression, around 15 MB for text recognition once its language model is counted, and far less for the rest. That load happens once per tool per session, and the runs that follow start immediately. The engines come from the site itself rather than from an outside network, with a single exception: the language model for text recognition, which is downloaded the first time you use a given language.

Does the site keep a history of what I process?

A history is kept, but it stays in your browser’s local storage and is never transmitted. Each tool remembers its last four runs, and a line holds only this: the name of the tool, the format of the file produced — an extension in capitals such as PDF, JPG or DOCX, followed by a count such as ×3 when there were several — the size of the result, and a timestamp. Your own file’s name does not appear there, and neither does any part of its content. You can wipe it from the interface, or by clearing the site’s data in your browser.

Which browser do I need?

Any recent version of Chrome, Edge, Firefox or Safari will do: the engines are compiled to WebAssembly, which all four have supported for years. Two caveats are worth knowing. Private browsing usually clears the cached engines when you close the window, so the loading time comes back. And only Safari decodes HEIC natively, which is why iPhone photos should go through the HEIC → JPG converter before any other image tool.

What should I do if a tool gives the wrong result?

Write to contact@nexavotools.com stating the tool, the browser, the operating system and the exact error message, without attaching the file concerned. The Contact page lists the details that help and the common causes, which usually come down to available memory or to private browsing mode.