Free, private PDF tools
Merge, split, zip & extract — nothing uploaded.
Most online PDF tools upload your document to their servers — fine for a meme, not for a resume, contract or ID scan. JobStraight's PDF Toolkit does everything in your browser instead: your file is read into memory, processed on your own device, and never sent anywhere. That's genuinely private, and it works offline once loaded.
Merge several PDFs into one, split out the exact pages you need, rotate or delete pages, turn photos into a PDF, stamp a watermark across every page, add page numbers, or pull all the text out of a PDF for your notes or an LLM — all free, no account, no watermark on your output. Powered by the open-source pdf-lib and pdf.js libraries running client-side.
PDF → ZIP handles the case splitting alone always left half-finished. Splitting a 40-page document gives you 40 files, and most tools then hand you 40 separate download links to click one at a time. This packages every page as its own PDF and gives you a single .zip. The archive is assembled in your browser too — the same privacy rule applies, nothing is uploaded to build it — and any multi-file result from Split or PDF→Images gets the same one-click ZIP button.
- ✓Merge, split, rotate, delete pages, images→PDF
- ✓PDF → ZIP: every page as its own file, one download
- ✓Watermark, page numbers, PDF→Text, flatten & sign
- ✓100% in your browser — files never uploaded
- ✓Free, no signup, no watermark on output
Why client-side processing protects your documents
Every major online PDF tool—ilovePDF, SmallPDF, Adobe's online tools—uploads your file to their servers. This happens invisibly: you click "choose file", select a PDF, and the tool processes it. The document leaves your device, reaches a server somewhere, is processed, and is returned to you. For documents containing sensitive information—a resume with your phone number and email, a contract with financial terms, an ID scan, a medical form, or a confidential project document—this upload is a security problem. The files are usually processed in memory and deleted after a few minutes, but the upload itself creates a window of exposure and trust.
Client-side processing means the file never leaves your device. Your browser reads the PDF into memory, processes it using JavaScript libraries (typically pdf-lib, pdf.js, or similar) running inside your browser, and outputs the result without ever sending the file over the network. The processing is slower than server-side (your device has less compute than a server), but for most PDF tasks—merging two documents, deleting a page, rotating an image, splitting a PDF—the speed difference is under one second. The privacy gain is absolute: no server knows what you uploaded, no logging occurs, no file passes through anyone else's infrastructure.
This is not theoretical. Privacy-focused users routinely avoid cloud-based PDF tools specifically because of upload concern. For a tool designed to handle resumes, contracts, and forms, client-side is non-negotiable. The tradeoff is that client-side tools cannot handle extremely large files (a 500 MB PDF might crash your browser) and they cannot parallelize work across a cluster. But for document splitting, merging, rotating, and text extraction—the most common PDF tasks—client-side is faster and absolutely private.
The technical challenges of splitting PDFs and why naive approaches fail
Splitting a PDF sounds simple: extract pages and write them to new files. The implementation is harder. PDFs are not a simple sequence of pages; they are a graph of objects—streams of compressed content, font dictionaries, images, annotations, and cross-references. Extracting a page requires understanding which objects belong to it, which are shared across pages, and which are global (like a font used throughout the document). A naive parser that treats each page independently often duplicates shared resources or breaks references, producing pages that look correct but are malformed.
Most naive splitting tools produce working output, but the resulting PDFs are bloated and fragile. A 10 MB PDF split into 10 pages should produce ten ~1 MB files. A naive splitter often produces ten 2-3 MB files because it duplicates the entire font table and image cache for every page. More problematically, some splits fail silently. A page with a scanned image embedded using JBIG2 compression can render in the original PDF but fail to render in the split file if the splitter does not preserve the compression settings. The page looks blank.
The right approach uses a PDF library that understands the specification, reconstructs object references correctly, and preserves stream compression. pdf-lib handles this properly in the browser. It reads the PDF, builds an internal model of the page tree, and when splitting, creates a new PDF with a subset of pages, updating all references and shared resources correctly. The output is valid and compact. The cost is memory—parsing a large PDF into memory is slower than streaming, which is why server-side tools can handle multi-gigabyte PDFs but client-side tools tap out around 100-200 MB on most devices.
Common PDF workflows and how to handle them without uploading
The most frequent need is extracting pages. A job seeker might have a thirty-page application package (cover letter, resume, portfolio samples) but a given employer only wants the resume and a specific project writeup (pages 5-7 and 15-20). Splitting out pages 1-3 and 5, 15-20 as separate PDFs, then re-merging into a custom submission package, saves time and ensures consistency. This workflow involves three operations: split the original PDF into individual pages, select the pages you want, and merge them back. Done on a server, this takes two minutes and three requests. Done in a client-side tool, it takes thirty seconds and zero uploads.
A second workflow is converting images to PDF—candidate photos, scanned documents, or screenshots. A recruiter might ask for additional materials as images ("send a photo of your work") and the candidate needs to combine them into a single PDF for submission. Client-side image-to-PDF tools read the images from your device, assemble them into a PDF (preserving aspect ratio, handling rotation, adding page breaks if needed), and download the result. This works offline and requires no upload. The quality and page layout depend on the tool's logic, but the fundamental operation—reading local images and assembling a PDF—is reliable.
A third workflow is text extraction, often underestimated. A candidate might need to pull all text from a PDF to feed into an LLM, to make a scanned document searchable, or to extract data for a form. Client-side extraction tools read the PDF and output plain text. For born-digital PDFs (created by exporting from Word or Google Docs), extraction is perfect. For scanned documents or PDFs with OCR disabled, there is no text layer, and extraction fails—the data is in the image, not the document. Client-side OCR is possible (Tesseract runs in the browser) but slow and resource-intensive. Most client-side tools skip it and report that extraction is not available for image PDFs, which is honest and prevents the worse failure of outputting garbage.
Frequently asked questions
Can I use your PDF tools offline?
Yes. Once the page loads, the PDF tools run entirely in your browser. If your internet disconnects, you can continue working. Your changes are processed locally. You need internet only to load the page initially and to download the final result—the processing itself is offline and private.
How large a PDF can I split or merge?
Depends on your device memory. Most browsers can handle 100-200 MB PDFs comfortably. Larger files may slow down or fail. Check your browser's memory usage before attempting very large files. Server-based tools handle gigabyte files; client-side tools trade file size for privacy. For typical documents under 50 MB, you will see no delay.
Does merging PDFs change the file size or quality?
Merging combines PDFs without recompressing. File size is roughly the sum of input files, sometimes slightly smaller if duplicate resources are detected and merged. Quality is preserved—no re-rendering or quality loss occurs. The output PDF is as readable as the inputs.
Can I add watermarks to every page at once?
Yes. Watermarking reads the PDF, overlays text or an image on each page at your chosen opacity and position, and saves the result. This works for stamping "CONFIDENTIAL" or "DRAFT" across all pages. Position and opacity are customisable. The output maintains the original content underneath the watermark.
What happens to my files after I download them?
They are yours. No copy is retained by our servers—client-side tools do not send files anywhere. Once downloaded, they are on your device. We have no access to them. Delete them from your Downloads folder whenever you choose, or use them immediately. Complete privacy and no tracking.
Part of Free utilities — 6 free tools in this set.
Part of JobStraight — the honest AI job-search platform. Runs in your browser; your data stays on your device.