Getting started

Steps

1. RunDouble-click PDFTextReplacer.exe. Python is not required. On first start the interface language follows the Windows display language; change it with "Interface language" in the bottom-right corner.
2. Choose a fileDrag a PDF or image anywhere into the window, or click "Choose file…". If the file is encrypted, a password box appears.
3. Set keywordsIn "Single" mode, fill in "Find text" and "Replace with" (leave it empty to delete). In "Batch" mode, import a keyword table.
4. FindClick "Find" in the bottom-left corner. Scanned pages are recognized starting from the current page. Results appear as recognition proceeds, with progress and estimated time remaining at the top.
5. ReviewClick a row and the preview on the right jumps to that page and marks the location. Click the first column to change ☑ to ☐, and that match will not be replaced.
6. Save asClick "Replace selected and save as…". After saving you can choose "Open result", "Open containing folder", or "Continue on result" to run another round of replacements.

Window layout

• Left panel: settings in steps 1-2-3, with a single main button at the bottom. The button label changes with the stage: Find → Replace selected and save as… → Open result.
• Middle panel: search results. The summary at the top lists the total count, keywords not found, entries merged because they overlap, and entries that need review.
• Right panel: page preview. You can switch it to “Preview replacement”.

Two work modes

“Work mode” at the top of the left panel switches between “Correct & replace” and “Secure redaction”. Redaction actually deletes information and rewrites the file. See Secure redaction.

Advanced layout options

Collapsed by default. The summary line shows the current values:

• Alignment: left, right (useful for right-aligned number columns such as amounts) or center.
• When image text is too long: allow overflow, compress (down to 60%), or stay within the original area.
• Loose matching: recognition ignores differences between 0/O, 1/l, 5/S and letter case. Matches found this way are labeled ”≈ loose”.

Message log

The status bar at the bottom of the window shows only the current status line. All messages are kept in time order in the “Message log”, where you can review and copy them. Error dialogs explain the cause and what to do next.

Settings are remembered

Recognition language, font, alignment, work mode, batch output folder, window size and similar settings are saved automatically. File contents, keywords and passwords are never written to settings.

FAQ

Is the original file changed?

No. The result is always saved as a new file.

Can I stop partway through?

Yes. While processing, the main button turns into a red "Stop". Results found so far are kept, and you can "Continue from page N". Stopping during a save does not leave a half-written file.