No, files are never sent to any server. All processing is done entirely in your browser. The library (pdf-lib) manipulates PDFs directly in the browser, so no network communication occurs. You can safely use this tool with confidential documents.

Roughly the same as the total size of the original PDFs. The merge process simply concatenates pages without modifying the content, so there is no quality loss.

However, the size may vary slightly depending on how fonts and resources are shared between files.

Password-protected PDFs cannot be processed. An error will be displayed when loading. Please remove the password using Adobe Acrobat or another tool before using this tool.

There is no strict limit, but it depends on your browser's memory. As a guideline:

- 10 files or fewer: No issues
- 10 to 50 files: Fine if total size is under 100MB
- 50+ files: Browser may become unstable if total size is large

For large numbers of files, consider merging in batches of a few files at a time.

In-page links and form fields are generally preserved. However, PDF bookmarks (outlines) may not be carried over during the merge. This is a limitation of the pdf-lib library.

Yes. You can add as many images, seals, and text stamps as you need. Each can be freely dragged and positioned on the page.

Transparent PNG is recommended. JPEG images will show a white background. You can also create stamps using the Seal Maker tool and upload them here.

Yes. Drag to move, use corner handles to resize, and adjust opacity and rotation in the properties panel.

Yes. Use the "Apply to" selector to choose current page, all pages, or a custom page range.

Yes, CJK characters are supported. When non-ASCII characters are detected, the Noto Sans JP font is automatically loaded and embedded. You can use Japanese text such as "社外秘" for your watermark.

Yes! You can choose between tile mode (repeating pattern across the entire page) or single placement mode. When tile mode is off, you can select from 9 positions using the position grid (top-left, center, bottom-right, etc.).

Recommended settings:
- Diagonal watermark: Rotation -45°, Size 60-80, Opacity 15-25%
- Bold & visible: Rotation -30°, Size 100-120, Opacity 20-30%
- Subtle: Rotation 0°, Size 40-50, Opacity 10-15%

Yes. The watermark is simply drawn as an additional text layer on the PDF, so text selection and copying from the original PDF remain fully functional. The search feature is also unaffected.

Using PDF editing tools (such as Adobe Acrobat Pro), it is technically possible to remove it. The watermark is added as an independent text element, so someone who understands the PDF structure could delete it.

If you need a truly tamper-proof watermark, consider rasterizing (converting to images) the PDF before applying the watermark.

No, text cannot be extracted from scanned image PDFs that do not contain text data. This tool uses PDF.js to extract text data embedded within the PDF.

For scanned PDFs, you will see a message saying "No text data on this page." To extract text from scanned PDFs, OCR (Optical Character Recognition) is needed, which is under consideration for future implementation.

Heading detection is a simple heuristic based on font size differences. Text that is 1.3x or more larger than the average font size on the page is treated as a heading (Markdown ###).

Accuracy depends on the structure of the original PDF. Well-structured PDFs will yield higher accuracy, while design-heavy PDFs (flyers, etc.) may produce false positives. If not needed, uncheck "Detect headings by font size."

Text is extracted, but the table structure (rows/columns) is not preserved. Text within tables is retrieved based on position information and may appear scattered, making it difficult to maintain the original table structure.

For accurate table data extraction, consider using specialized tools like Adobe Acrobat or Tabula. Chart data is often rendered as images and is not subject to text extraction.

Choose based on your use case:

Markdown (.md) is recommended when:
- Pasting into Notion, GitHub, etc.
- You want to preserve heading structure
- You plan to edit and format the text later

Text (.txt) is recommended when:
- Copying into an email body
- Importing into other software
- You only need simple plain text

PDF.js supports multiple languages, but vertical text PDFs may have disrupted reading order. Horizontal text PDFs are extracted most accurately.

Right-to-left languages such as Arabic and Hebrew are generally supported, but complex layouts may result in disordered text.

Yes, all file formats are supported. You can rename and bundle images, Excel, Word, video, audio, or any other type of file into a ZIP.

Files are numbered in the order they were dropped or selected. If you want to change the order, drag and drop items within the file list to rearrange them.

No. Renamed files are downloaded as a new ZIP file only. No changes are made to the original files.

The original extensions are preserved. For example, with prefix "doc_": report.pdf becomes doc_001.pdf, data.xlsx becomes doc_002.xlsx — the extension follows the original file.