Browser-only tools: PDF to Excel and image compression without uploads

I built two online tools that share one property: your file is never uploaded. Jadvalsaz converts tables in PDFs to Excel and the image tool compresses or converts images. Both run entirely in the browser.

Why no server?

Privacy: a PDF may be a report card or a stock sheet; if it never leaves the device you do not have to trust anyone. Cost: no processing server, the site is static. Speed: no upload or download.

How Jadvalsaz builds a table from a PDF

A PDF has no "table", only pieces of text with coordinates. pdf.js gives each piece's (x, y) position and the rest is algorithm: cluster pieces by Y into rows; build columns by clustering the right edge of pieces (Persian text is right-aligned) with an adaptive tolerance; merge fragmentary columns and drop empty ones; write an xlsx with SheetJS, one sheet per PDF page, right-to-left view.

Limits: 20 MB and 50 pages. With no machine learning involved, irregular tables may be detected badly, and scanned PDFs (images with no text) are not supported.

Image compression with Canvas

The browser can draw an image onto a canvas and export it again at a chosen quality. The tool does exactly that with a quality slider, optional resizing and a target format; changing the quality reprocesses the image.

A practical number from this very site: after converting to WebP the founder photo is about 13 KB and the NØX hero image is 19 to 69 KB across three sizes.

Costs of this approach

Libraries load from a CDN at fixed versions, so internet is needed; big files on weak devices can be slow; and I have no usage statistics because nothing reaches me.

It suits light processing of sensitive files. Heavy work (OCR, large models) needs a server, and then you should tell users honestly that the file leaves their device.