PDFbuff Workspace — Private PDF Editor and Tools

Clear and apply to leave a blank in place. Other text stays put. Visual covering only, not secure redaction.

Font Mode

Exact Fidelity Image Approximation(visual approximation) Upload Original Font (TTF / OTF)

Select or drag in multiple PDFs to merge in list order.

📂 Choose PDF Files

Split the current PDF. Mode: page range (e.g.1-3,5,7-9) or fixed page count (N pages per file).

By page range By fixed page count

Rotate the current PDF. Angle: 90°/180°/270°, for all pages or selected pages.

All pages Selected pages

Delete selected pages from the current PDF (e.g.2,5,8-10）。

Extract selected pages from the current PDF into a new file (e.g.1,3,5-7）。

First create a signature image (handwrite or upload a transparent PNG), then drag, place, and scale it on the PDF page, and finally confirm to embed and save.

⚠️ Capability Limits: this feature embeds a handwritten or uploaded PNG image into the PDF,it is not a digital signature; it does not provide certificates, identity verification, trusted timestamps, or audit trails. Modifying the PDF may invalidate any existing digital signatures.

🖼️ Upload Signature (PNG only)

Choose an image (JPG/PNG); in placement mode, click on the page to place it. Drag to move, use corner handles to scale, the top handle to rotate, and click ✕ to delete.

Capability limits: the image is a new overlay element (original content is kept); it does not replace or delete the original image in the PDF, and the exported size is not guaranteed to be smaller.

Choose Image (JPG/PNG)

Auto-detect form fields in the current PDF (text/checkbox). After clicking "Start Filling", fillable controls appear on the page; on save, values are written into the PDF, which remains an interactive form.

Capability limits: only supports documents withalready present AcroForm fields (text, checkbox); creating fields in blank areas is not supported. The appearance of Chinese/complex fonts may depend on the reader, but the field values themselves are always preserved correctly. Dropdowns/radio buttons are best-effort.

Select or drag in multiple images (onlyJPG/PNG) to merge into one PDF in list order, one image per page.

📂 Choose Images

Page Size Orientation Margin pt

Export the current PDF's pages as images. Pages (e.g.1-5,8, blank = all), format PNG/JPG, with adjustable resolution and quality.

Format Resolution JPG Quality90

Compress the current PDF. Type is auto-detected and an initial mode is recommended,final result depends on what is generated：

① Structure Rewrite(text-based) — re-serialize while keeping the text layer and editability,size reduction not guaranteed；

② Image Re-encode(scanned/image-based) — re-encode each page as JPEG,loses the text layer and other PDF structure, suitable only for image-only PDFs.

Only when the candidate file isstrictly smaller will it be downloaded; if larger or the same, it is not downloaded and a notice is shown.

Mode JPG Quality70

Select multiple PDFs to run compression, image conversion, watermarking, splitting, rotation, page deletion, page extraction, re-encode compression, or batch signing.Batch processing is a Pro feature; the free version supports single-file processing only.

Choose PDF files (multiple allowed)

Operation

JPG Quality75

Recognize scanned / image-based PDFs intosearchable, copyable text PDFs. The original page images are fully preserved with a transparent text layer overlay (offline tesseract.js recognition; files never leave your device).

Language: Chinese (chi_sim) accuracy depends on training data — clearest, straightest scans work best; English (eng) is highly accurate.

Recognition Language Resolution (DPI)300

Minimum 300 DPI: measured Chinese recognition is extremely resolution-sensitive — at 200 DPI the character error rate reaches 54% (text layer effectively unusable), dropping to 8.5% at 300 DPI. Below 300 silently produces a PDF that looks successful but has a completely wrong text layer, so it is not allowed.

Export the current PDF to .docx — pick one of two modes.

Export mode

Editable mode: keeps the text layer and images, with font size, bold, italic, font, and page breaks — the content can be edited right after export. Honest limits: the layout may shift, and vector graphics and borderless tables are not reconstructed; for scanned files run OCR first.

Layout-fidelity mode: every page is rendered as one image filling the page, visually identical to the source PDF (including backgrounds, vector graphics, and table borders). Text is not editable or searchable — best for printing and archiving; the more pages, the larger the file.

Aggregate the text in the current PDF byrow/column coordinates into an .xlsx table (one worksheet per page).

Honest limits: only works fortext-based text-based PDFs; scanned files with no text layer yield empty content. Recognition of complex nested tables is limited.