About Deskbench

Deskbench is 41 free tools for the ordinary document work that fills a working week: merging a PDF before sending it, resizing a photo that is too large, converting a spreadsheet someone exported wrong, comparing two drafts of the same paragraph.

The premise

Every free tool site works the same way. You upload a file. A server processes it. You download the result — after an advertisement, sometimes after an email, occasionally with a watermark stamped across it. The upload is not a technical requirement. It is the business model. Your file is the price of admission, and what happens to it afterwards is governed by a retention policy you did not read.

None of it is necessary. Browsers have been capable of this work for years. pdf-lib can rebuild a PDF's page tree. The canvas API can resample an image better than most desktop software did a decade ago. SheetJS reads an xlsx file. All of it compiles to JavaScript and WebAssembly and runs in the tab you already have open.

What that means in practice

When you drop a file onto a tool here, it is read into the tab's memory using the browser's File API. It is transformed by library code running on your processor. The result is offered back as a download through a blob URL. At no point is there a network request containing your data, because there is no server to send it to.

You do not have to take our word for it. Open your browser's developer tools, switch to the network panel, and use any tool on this site. You will see the page load, the fonts load, the libraries load — and then nothing, no matter how large the file you process. Or simply disconnect from the internet after a tool page has loaded. It keeps working.

What browsers cannot do

This approach has real limits, and pretending otherwise would be dishonest.

Memory. A browser tab has a bounded heap. A two-gigabyte PDF will fail here and succeed on a server. Most work is well within the ceiling; some is not.

Compute-heavy tasks. High-quality background removal needs a segmentation model larger than anyone should download to crop one photo. OCR needs a trained recogniser and produces mediocre results at browser-feasible sizes. Video transcoding is out of the question. These are absent from this site rather than present in a poor version — a decision we have already made once, when background removal was considered and dropped.

Format support. Whether a HEIC file opens depends on whether your browser ships a HEIC decoder. We cannot supply one. When a file will not load, that is your browser telling you something true.

What comes next

More tools, chosen the same way: does this belong in a browser, and can it be done properly there? If the answer to either is no, it does not get built.

Found a bug, or a tool that produced output you did not expect? Tell us. Read the privacy policy, which is short because there is little to say.