Twelve focused tools and a fill-and-sign editor, all running inside your browser tab. No account, no upload, no export limit. The last section on this page is the list of things it cannot do, because knowing that up front is worth more than one more bullet point.
Each one does a single job and shows you the result before anything downloads. All 10 are also reachable from inside the editor, so you never have to leave a document to run one.
These live in the editor rather than on pages of their own, because they are things you do to a document you already have open.
Build a PDF from a blank page
The creation zone starts from nothing rather than from a file you already have: paper size, orientation and per-edge margins, headings, paragraphs and lists, local TTF/OTF embedding with explicit variants, images, measured tables, reusable headers and footers, vector rectangles, ellipses, lines and arrows, page controls, and eight starting templates. It exports selectable text, and the on-screen preview is generated from the same layout model as the downloaded file.
Fill any form, detected or not
A local scan reads three signals at once — the form fields already embedded in the PDF, the boxes and underlines drawn in the page geometry, and the labels in the text — and merges them into editable fields. Forms that arrive part-completed import their existing values, so you review rather than retype. Anything the scan misses you place by hand, and manual fields behave identically to detected ones.
Sign with a real drawn signature
Draw with a mouse, finger or stylus. Newly drawn signatures are represented as normalized vector strokes, so they have no background and stay sharp when resized. Legacy transparent bitmap signatures remain supported for documents and saved sessions created before the vector format was introduced.
Check and cross marks
Toggle a detected checkbox with a single click, or place a cross for the forms that expect an X rather than a tick.
Text-aware highlighting
Drag across a passage and one marker stroke snaps to each line it covers, blending like a real highlighter rather than sitting as a flat box on top. A single click marks a whole visible line.
Erasing that hugs the words
Drag the erase tool over text and the cover box follows the words underneath instead of guessing a rectangle. Match the colour to the paper when the page is not pure white.
Precise placement
Numeric X/Y controls and corner handles for the rows that are too tight to eyeball, plus zoom that snaps back to a full-page fit in one click.
Privacy is a property of the design
Documents are opened, scanned, edited and exported by your own browser. Nothing is sent to a server, and unsaved work disappears when the tab closes. The explicit Save on this device action is the exception: it writes a local IndexedDB session only after you ask, never stores a password, and keeps the session on that device until you delete it or the browser clears site data. You can verify local processing rather than take it on trust: install the site, switch on airplane mode, and run an export. Work that finishes with the network off was never being sent anywhere.
The site is free and funded by advertising. Advertising and analytics are consent-gated, and neither ever sees your document — they cannot, because your document never leaves the tab.
What it does not do
Every one of these is a real limitation rather than a feature that is “coming soon”.
Cover is not redaction; Remove has trade-offs
Cover paints over the selected area, so underlying content can still be recovered. Remove rasterizes each affected page so the covered content is absent from the exported page, but that also removes selectable text, links, form fields, annotations, and tags on that page. This is not a certified redaction workflow: inspect the export before sharing it.
OCR is local, language-limited, and reviewable
A scanned page starts with no machine-readable text, so field detection and text-aware highlighting cannot use it directly. The OCR tool can add a searchable overlay in English, French, or Spanish using local language assets; review low-confidence results before export. OCR does not automatically create form fields or add a tagged-PDF structure tree.
Conversion, compression, and flattening are focused
PDF-to-images exports PNG or JPEG files and bundles multiple pages in a ZIP. Images-to-PDF accepts PNG and JPEG files and makes one page per image. Compress offers lossless cleanup or a smaller raster result; raster compression removes selectable text, links, forms, annotations, and tags. Flatten bakes AcroForm values only, not every annotation subtype. Word-to-PDF is not available.
Permission flags depend on the reader
Protect applies real AES-256 encryption, but printing and copying restrictions are requests that the receiving PDF viewer chooses to honour. Encryption is enforced; permissions are advisory.
Very large documents are bounded by your device
Everything runs in your browser tab, so a several-hundred-megabyte file is limited by your machine's memory rather than by a server queue. That is the direct trade for nothing being uploaded.
More detail on the ones that catch people out is in the guides and the FAQ.