Help — Copy Diff Tool

🚀 How to use it

Load Document 1 and Document 2 — upload a file or switch the Input Mode to Paste Text.
Click "🔍 Compare Documents" — a side-by-side diff appears below.
Read the results — red strikethrough means removed, green means added.

💡 For scanned pages or images, use the ✂️ Crop button to select a region and run OCR on just that area before comparing.

📄 PDFs and PPTX files show a Pages to compare field after loading. Leave it empty to use all pages, or enter a range like 1-3, 5, 8.

🔦
Live HighlightMarks extracted text on the document so you can check coverage before comparing.
✂️
Crop ToolDraw a box around any region and OCR just that area.
🗂️
Crop QueueQueue multiple regions across pages, then extract them all at once.
📑
Page rangesCompare only the pages you need from PDFs and presentations.
🔍
Diff searchSearch inside the comparison results and step through matches.
🔤
Word or character diffSwitch granularity in the results toolbar.
🌙
Dark modeYour preference is saved and shared with this help page.
🔒
100% localNo files or text are ever sent to a server.

Click ✂️ Crop on a loaded document.
Drag to draw a selection, then click ➕ Add to Queue. The canvas resets so you can draw another.
Navigate pages or switch documents — keep adding regions.
When done, click 📥 Use as Doc 1 or 📥 Use as Doc 2. All queued regions are OCR'd and combined into that document slot.
Click 🔍 Compare Documents as normal.

💡 Use the 📄 Extract Text (OCR) button for a quick single-region extraction without queuing.

Check the file format is supported (PDF, DOCX, PPTX, JPG, PNG, TIFF, WebP, HTML, TXT). Password-protected PDFs cannot be opened — remove the password first. Very large files may take a few seconds; watch the loading indicator.
Higher resolution images give much better results — 300 DPI or above is recommended. Use the Crop tool to isolate a clean text area rather than running OCR on a busy full page. The language model downloads once on first use; wait for it to fully load before extracting.
The tool detects multi-column layouts automatically, but very complex mixed layouts may need manual help. Use the Crop tool to select individual columns or sections and queue them in the correct reading order.
Toggle 🔦 Live Highlight to visually verify that both documents extracted the text you expected. Try switching the granularity from Characters to Words for a cleaner result. If using page ranges, make sure both documents cover equivalent content.
Use page range filtering to process only the pages you need. Close other browser tabs to free memory for OCR. For large PDFs, consider splitting the comparison into smaller page ranges across multiple sessions.