How QuickToolio processes your files in the browser

When you select a file in a QuickToolio tool, your browser reads it directly from your device. There's no upload step in between: the file never travels to one of our servers. The tool transforms it using the browser itself as the computing engine, the same way an installed program would use your machine's processor.

Selecting a file (via the system picker or by dragging it) isn't the same as uploading it: selecting only gives the browser permission to read it into memory; uploading would mean sending it over the network to a server, which no QuickToolio conversion, compression or editing tool does.

The result — the merged PDF, the compressed image, the extracted audio — is also generated on your device, and offered for download straight from there, without passing through any server in between.

The foundation is JavaScript, run by the browser itself, together with the APIs modern browsers expose for working with files, images and audio: the File API, Canvas for drawing and manipulating images, the Web Audio API, among others.

For the more demanding operations, several tools use WebAssembly (WASM), a format that runs code at near-native speed inside the browser. For example: removing an image's background or increasing its resolution uses an AI model running on ONNX Runtime over WebAssembly; opening a password-protected RAR uses a WebAssembly library equivalent to a desktop archive tool's; recognizing text in a scanned PDF (OCR) uses Tesseract, also compiled to WebAssembly.

The video and audio tools use FFmpeg compiled to WebAssembly, running inside a Web Worker: a separate thread that does the heavy work without blocking the rest of the page while you convert or trim a file.

By contrast, PDF, Word (DOCX), OpenDocument (ODT) or ZIP tools run on plain JavaScript, without WebAssembly, because those particular operations don't need that extra performance.

Even though your files are processed locally, loading the page itself does involve network traffic, like on any website: the HTML, CSS, JavaScript and, when a tool needs it, the WebAssembly files or AI models mentioned above. All of that is the site's own code, never your file.

QuickToolio is served through Cloudflare, which acts as our hosting and content-delivery provider: it technically handles delivering the page, as it would for any website. The homepage's smart search is the only feature on the site that queries a cloud service (a Cloudflare Worker with AI): it sends only the text you type into the search box and the language, never a file or the content of any tool.

Analytics (Google Analytics) and advertising (Google AdSense) both work under the same rule: they only load when enabled in the site's configuration and you've given explicit consent; without both conditions, no analytics data loads or is sent, and no ad is shown. When analytics is active, it measures which tools get used in aggregate, never the content you enter into them.

Your browser stores, on your own device rather than a server, two very specific preferences: your choice about analytics and advertising, and whether you prefer the light or dark interface theme. Neither one contains your file or any data about what you process in a tool.

• Privacy: your file's content doesn't pass through anyone else's hands to be processed, because it never leaves your device.
• No upload to wait for: there's no need to wait for a large file to finish uploading to a server before you can start working with it.
• You can start as soon as the tool loads: once your browser has downloaded what that specific tool needs, it can process without any further requests to a server.

This doesn't mean QuickToolio works offline: you still need to load the page (and, the first time you use a tool, its specific files) before you can process anything.

Local processing has a trade-off: the limit on what you can do depends on your own device, not on a server with more resources. Available memory, processor power, the browser you use and its compatibility, and the file's size or complexity all affect whether an operation finishes fine, takes longer than expected, or fails.

Some tools (the ones using WebAssembly, like video, audio, OCR or AI-based image editing) need to download their engine the first time you use them, which adds some initial loading time. On very large files, or ones with many pages, images or tracks, an operation can take considerably longer than with a small file, and on devices with little memory it may close before finishing. There's no universal "maximum size" figure: it depends on the specific combination of file, operation and device.

You don't need to know how to code to get a sense of this. One simple signal: if you disconnect your Internet connection right after the page and the tool have fully loaded, and you can still process your file, that's a good sign that particular processing happens on your device. This doesn't prove there's no traffic at other points during loading, just that processing the file itself doesn't generate any.

If you're on a desktop browser, the "Network" tab in developer tools (usually opened with F12) shows the page's requests. You'll see requests when the tool loads; what you shouldn't see when you click "process" or "convert" is a new request carrying your file's weight. This is indicative, not absolute proof: the page's own loading, or auxiliary features like the smart search, generate their own requests, so not every request you see is related to your file.

While you work, the file you select and the result the tool generates live in your browser's memory, not in any storage of ours. If you close the tab, reload the page, or navigate to another tool, that memory is released: the original file and the result stop being available there, and you'd need to repeat the operation.

The only things that survive closing or reloading the page are the two preferences stored in your browser (theme and consent) and, of course, any file you explicitly downloaded to your device during the session.

Privacy policy

Cookie policy

Third-party licenses

Contact