Free · unlimited · open source · made in India
Merge, split, compress, convert, sign, scan and redact PDFs on your own machine. Nothing you open here is uploaded — not to my server, not to anyone else's. There is no server.
/vendor — no analytics, no trackers, no accounts, no upload endpoint to hit.
Join any number of PDFs into one. Page sizes, orientation and rotation preserved exactly.
DividePull out a range, break every N pages, or explode one file into many. Good for scanned invoice batches.
ArrangeSee every page, then reorder, remove, duplicate or turn them on a visual grid.
TurnFix sideways pages and save it into the file, so it stays fixed for everyone who opens it.
TrimCut wide margins so text fills an e-reader screen, with a live preview.
Downsamples the images inside and leaves the text as text. Tells you exactly what it changed and what it refused to touch.
ConvertPages as images at whatever resolution you actually need — 72 dpi for a screen, 300 for print.
ConvertPhotos and screenshots into one document. No watermark, no page limit.
CapturePhotograph paper with your phone camera. No app installed, and the photographs never leave the device.
Destroys what is underneath rather than covering it. A black rectangle is not redaction.
SignDraw a signature and place it. Your signature is as sensitive as the document — neither one leaves your device.
StampText or a logo across the pages, with opacity, angle and page range.
NumberPosition, starting number and wording. Skip the cover page properly.
Load a document once and stack the jobs — rotate, sign, number, compress — then download at the end. Nothing is written to disk along the way.
Many at onceOne design × many names. Paste a spreadsheet, drop your artwork, get every personalised certificate, badge or ID card as one PDF or a ZIP.
RoadmapEdit, OCR, form filling and more are being built. Office conversions and AI never will be, and the page explains exactly why.
Every other free PDF site works the same way: you upload your file to a company's servers, they process it there, and you download the result. That is fine for a restaurant menu. It is not fine for a signed contract, a medical report, a bank statement, or a scan of somebody's Aadhaar card — and those are exactly the documents people put through these tools every day.
Kagaz does the work in your browser instead. That single decision has some pleasant side effects: there are no daily task limits, because running the job costs me nothing. There are no file size caps beyond what your own machine can hold. There are no accounts. And once the page has loaded, you can turn off your WiFi and keep working.
Being honest about the edges is the whole point, so: there is no PDF to Word conversion here, and there never will be. Doing it properly means reconstructing a layout from a page description, and it cannot be done well inside a browser. Every free tool that offers it is sending your document to a server. If you need it, that trade-off might be worth it to you — but you should know you are making it.
The same goes for office documents in the other direction, for turning a web page into a PDF, and for anything involving AI. The roadmap lists every one of them with the reason.
Kagaz also will not remove a password you do not have, and will not claim that a signature drawn on a page is a legally binding e-signature. When compression cannot safely shrink an image — a CMYK scan, a JPEG 2000 photo — it leaves it alone and says so, rather than corrupting it to make a number look better.
You will not find a “2 million documents processed” badge anywhere on this site, and there never will be one. A number like that has to be counted somewhere, which means a request fired from your browser every time you open a page or press a button. It would be tiny, and anonymous, and it would still be this site watching you.
So I have no idea how many people use Kagaz, because finding out would mean watching you. That is the trade, and I would rather have it this way round.
This is not a promise you have to take on faith. The repository's test suite fails the build if any page ever loads something from another origin, or if the shared scripts ever gain a network call. If you want a number that tells you whether this is worth trusting, the repository is public and so is its star count.
Plain HTML, CSS and JavaScript. No framework, no build step, no bundler. Each tool is one self-contained file. The heavy lifting is done by pdf-lib, pdf.js and JSZip, all vendored into the repository rather than pulled from a CDN — a CDN would quietly break the offline promise.
If you want to change something, clone the repository and open a file in your browser. That is the entire development environment.
I'm Amit Jena — an Accidental Designer. Mathematics and computer science by training, pulled sideways into design, and now mostly building data-heavy things: dashboards, document tools, anything that turns a pile of numbers or paper into something a person can use.
Nearly all of it is open source, and I'd rather it stayed that way. Partly because a tool asking for your trust should let you check it, and partly because I have never learned more from anything than from reading other people's code.