Skip to content

The document viewer

The left side of the editor is the document itself, rendered so you can read it the way a recipient would. Every format renders on your machine — opening a document to view it sends nothing anywhere.

PDFs render page by page with a real, selectable text layer on top. That text layer is what lets you:

Scanned PDFs with no text layer still render as images; the model read them through vision during the run, so the entry list is populated even when there’s no text to select.

A page indicator floats at the bottom so you always know where you are in a long document.

Word files render as a reflowed reading view — the text, headings, lists, and tables in document order, styled for reading on screen. This is for reviewing content, not layout.

The saved .docx keeps its original layout, fonts, and formatting. The on-screen preview is a clean reading copy, not a pixel match. A status-bar note reminds you of this whenever a Word file is open.

In Word and text files the drawing tool becomes a highlighter: select text and redact it, instead of dragging coordinate boxes.

Shown as-is. Markdown is rendered as source so you can see exactly what will be written back — links and headings are preserved untouched in the output.

The viewer can show either the anonymized result or the original input, and can overlay an instant-redact preview. Those switches live in the status bar — see reading what changed.

Click the file badge in the top-left for the origin file and output file paths, plus size and page or image counts. Handy when you have several editors open and need to confirm which document is which.

If the document carries images the model flagged — a signed scan, an ID photo, a logo — they appear as thumbnails in the entries panel with a blur switch each. See redacting images.