Document Editor — Edit Scanned PDFs & Documents Online

How the editor turns a scan into editable text

When you upload a photo, a scan or a PDF, the page is cleaned up first: in a phone photo the document’s edges are found and the perspective flattened, the page is straightened, uneven shadows are lifted and light noise is removed, all at 200 DPI. Then the recognizer reads every line on every page and records what it says, exactly where it sits, and how confident it was. What you get back is your real page — the original image — with an editable block over each paragraph, not a re-typed approximation of it.

Your edits are kept separately from that reading pass, and that separation is what makes the result look right on the way out. Exporting replays only what you changed: edited and deleted lines are erased and repainted, while every part of the page you never touched keeps its original pixels — the same paper texture, logo, stamps and signature. Blocks the recognizer was unsure about are marked so you can check them before exporting, and every block stays fully editable whether it was flagged or not.

What you can change on the page

Text first. Click a paragraph to rewrite it, double-click a single line or word to fix just that much, or select across characters to cut, paste over or delete part of a sentence. Blocks move, resize and rotate; font size, spacing and alignment live in the properties panel; and the editor can measure a block’s existing font and offer it to you with a confidence score, which is applied only when you accept it — nothing changes your document behind your back. You can also add new text where the original had none. Find and replace works across the whole document, and spell checking runs against real dictionaries (English, Spanish, French, Dutch and Portuguese) with a personal word list for names it should stop flagging.

Beyond text there are black-box redactions, highlights, rectangles, ellipses, lines, arrows, polygons, and a signature you draw yourself. Everything you add is an object you can restyle, duplicate, reorder, lock or hide by layer. Undo and redo cover every change, versions are snapshotted as you work so you can jump back to an earlier state, and pressing ? lists every keyboard shortcut — a list generated from the editor’s own key table, so it can never advertise a shortcut that does nothing. Touch works as well as a mouse, so a document can be fixed on a phone.

What it handles well — and what it doesn’t

Clear scans and well-lit phone photos of printed documents are what this is built for: invoices, contracts, forms, letters, receipts, certificates, and pages from books or reports. Recognition ships English models today. Handwriting, heavy blur, glare and very low resolution are genuinely hard, and text pulled out of them can come back wrong — the tool marks that uncertainty instead of hiding it, and any block can simply be retyped.

You work on one document at a time: a JPG, PNG or PDF up to 50 MB and 20 pages. A longer PDF is refused up front, with its real page count, rather than half-processed. Reading a document takes a while — OCR on a long file is measured in minutes rather than seconds — which is time spent once, before an editing session that runs entirely in your browser.

Exporting to PDF, PNG or JPG

Every export is rendered from the page images, which has two consequences worth knowing before you download. The first is fidelity: the PDF you get is a picture of each page rather than selectable text, and that is exactly why it looks like the document you edited — same layout, same fonts, same everything you left alone. The second is that a redaction is final: the black box is flattened into the pixels, so there is nothing underneath for anyone to move aside or copy out.

PDF pages are lossless by default. If the file has to be smaller, switch on JPEG compression in the export dialog and pick a quality, or run the finished download through the PDF compressor; our guide to compressing a PDF explains which of the two to reach for and why a scanned PDF shrinks so much more than a text one.

Where your document goes

Reading and exporting run on our own servers over an encrypted connection. Recognition happens on an internal service that is not reachable from the internet, and your file is never handed to a third-party AI provider. The editing itself never leaves your browser: typing, redacting and drawing are local, and nothing is uploaded again until you press Export. Your work autosaves in the browser so a refresh does not lose it, and both the uploaded document and the exported result are deleted automatically after 24 hours. No account, no watermark, free.