Browser PDF Editor

The Browser PDF Editor is a complete PDF workstation that runs entirely inside your web browser. It is not a single-task utility that merges two files or stamps a signature — it is one tool that covers the whole lifecycle of a PDF: viewing, page organization, annotation, true text editing, merging, splitting, compression, format conversion, OCR, form filling, and signing.

Under the hood, pages are rendered with PDF.js while every document mutation is handled by a custom Rust engine compiled to WebAssembly, running in a web worker so the interface never freezes. There is no account to create, no watermark on output, no file size quota, and — most importantly — no upload. The PDF you open stays on your device from the first click to the final export.

Most free PDF sites work by uploading your document to their servers, processing it there, and sending back a result. That model has two costs: you wait on two network transfers, and your file — which may be a contract, a medical record, or a payslip — sits on someone else's infrastructure under whatever retention policy they choose.

This editor takes the opposite approach. The processing engine is written in Rust and compiled to WebAssembly, which runs at near-native speed inside your browser. You get the performance of a desktop application with the privacy of local processing: a 200-page merge or an Extreme-level compression completes in seconds on your own hardware, and the document never crosses the network. Because the heavy work happens in a web worker, the viewer stays responsive while the engine crunches.

The editor is organized around eight capability areas:

• Pages: drag thumbnails to reorder; rotate, duplicate, extract, and delete pages with a click.
• Annotation: pen, marker, highlighter, shapes, arrows, sticky notes, text boxes, and images placed anywhere on a page.
• True text editing: click any text line and rewrite it in place — the engine edits the actual PDF text blocks. Find & replace works across the whole document, and designer header/footer styles propagate to every page automatically.
• Document operations: merge multiple PDFs into one, split a PDF by page ranges, extract ranges into a new file, and compress with four quality levels from Low to Extreme.
• Convert: export pages as JPG, PNG, or WebP images, export the text as .txt or .html, or go the other way with images to PDF and text/HTML to PDF.
• OCR: make scanned pages searchable in English, Hindi, or Gujarati — entirely on your device.
• Forms & signing: fill text fields, checkboxes, radio buttons, and dropdowns; create new fields; sign by drawing, typing, or uploading a signature image.
• Batch mode: process many PDFs in one run and download everything as a ZIP.

Most free tools "edit" text by drawing a white box over the old words and placing new text on top. Zoom in, print, or copy the text and the trick shows. This editor is different: when you click a text line and retype it, the Rust engine rewrites the underlying PDF text blocks, so the change is part of the document itself — selectable, searchable, and clean at any zoom level.

The same engine powers find & replace across the whole document — useful for fixing a misspelled client name in a 40-page proposal — and a set of designer header/footer styles that, once applied, propagate to every page automatically. Update the header once and the engine keeps all pages in sync.

Here is how the pieces combine in practice:

• Assemble a report from several PDFs: drop the source files in, merge them, drag thumbnails to fix the page order, delete the duplicate cover pages, apply a header with your document title, and export one polished file.
• Make a scanned contract searchable, then sign it: run OCR on the scan so every clause is findable with full-text search, use the search to jump to the signature block, sign by drawing or uploading your signature, and fill any remaining form fields.
• Compress for email: a 25 MB scan won't clear most attachment limits — pick a compression level, preview the result, and step up to High or Extreme only if you need to. Batch mode compresses a whole folder of PDFs in one run.
• Redline a draft: highlight problem passages, attach sticky notes with your comments, strike sections with the pen, and drop text boxes where wording should change — then export a marked-up copy for the author.

The editor opens standard (non-encrypted) PDF files of any page count. For conversion, it exports pages as JPG, PNG, or WebP at 72, 144, or 216 dpi — either the current page or all pages bundled into a ZIP — and exports document text as .txt or .html.

Going the other direction, it builds PDFs from JPG, PNG, and WebP images directly, rasterizes SVG, AVIF, GIF, and BMP on the way in, and converts plain text or HTML into paginated PDF documents. If you only need one of these conversions, the dedicated PDF to text extractor and other single-task pages do the same job with a simpler interface.

Three design decisions keep the editor fast. First, the mutation engine is compiled Rust, so operations like merging, splitting, and compressing run at near-native speed rather than interpreted-JavaScript speed. Second, the viewer uses virtualized rendering: only the pages near your viewport are drawn, so a 500-plus-page manual or a very large scanned book scrolls as smoothly as a two-page letter. Third, batch mode runs on its own worker, separate from the editor's engine worker, so you can keep reading and annotating one document while a stack of others is being compressed in the background.

Full-text search with highlighting, single-page and continuous scroll modes, dark mode, keyboard shortcuts, and touch support round out the experience on both desktop and mobile.

Every operation — rendering, editing, OCR, compression, export — executes in your browser's own sandbox. No file is ever uploaded, there is no server-side processing, and no account or email is required. Close the tab and the document is gone from memory. Even OCR, which most tools outsource to a cloud API, runs on-device via Tesseract compiled to WebAssembly.

The only data the tool persists lives in your browser's IndexedDB, on your machine: saved signatures (so you don't redraw them every session) and the session-recovery autosave, which snapshots your working file and edits (for files up to 80 MB) so an accidental tab close doesn't lose your work. Nothing in IndexedDB is transmitted anywhere. To wipe it, use the tool's own clear/reset options or clear site data for tool-web.dev in your browser settings — standard browser storage controls apply.

A few things this editor deliberately does not do:

• Encrypted PDFs are not supported. Password-protected files must be unlocked with the tool that created them before opening here; the editor does not attempt to bypass or crack protection.
• OCR covers three languages — English, Hindi, and Gujarati. Scans in other languages will render fine but won't gain a searchable text layer.
• Signatures are visual, not cryptographic. Drawing or uploading a signature places an electronic signature image in the document, which is widely accepted for everyday agreements, but it is not a certificate-based digital signature with cryptographic identity verification.
• Session autosave caps at 80 MB. Larger files open and edit normally; they simply are not snapshotted for recovery, so export before closing the tab.

A few habits get better results:

• OCR before you search or extract text from a scanned document — without a text layer, search has nothing to find.
• Reorder and delete pages before merging additional files, so the thumbnail panel stays easy to reason about.
• Start compression at Medium. Step up to High or Extreme only if the file is still too large; heavier levels trade image quality for size.
• Lean on undo. With a 200-step history, it is faster to experiment and undo than to hesitate — and session recovery has your back if the tab closes.
• Choose 216 dpi for print, 72 dpi for the web when exporting pages as images; 144 dpi is a good screen-sharing middle ground.
• Export early and often for files over 80 MB, since those fall outside the autosave safety net.