PDF Toolkit Pro

PDF Toolkit Pro

The full PDF toolbox powered by PyMuPDF, running locally in your browser via Python (Pyodide). Real text extraction with coordinates, real redaction that removes the underlying text, embedded image export, annotation parsing, metadata editing, page reordering, merge and split. Nothing uploaded.

100% local PyMuPDF inside warming up...

Drop a PDF or click to browse processed locally in your browser, nothing uploaded

Drag to reorder. Click X to remove. Click a card to select.

Drop a PDF above to see page thumbnails.

Mode Plain text Blocks with positions Words with positions Full dict (JSON) HTML

Click Extract text to pull all text out of the PDF using PyMuPDF. Modes give you progressively more structure: plain text, blocks (paragraphs with bounding boxes), words (one row per word with positions), dict (full structured tree), or HTML.

Find (one phrase per line, case sensitive)

Options Case insensitive Treat as regex

Real redaction: text is removed and replaced with a black box.

Pulls every embedded image at original resolution (not page renderings).

No images extracted yet.

Flatten burns annotations into the page so they cannot be edited.

No annotations listed yet.

Title

Author

Subject

Keywords

Creator

Producer

Drag rows to reorder, click X to remove.

Add files to merge.

Ready.

What makes this different

Most online PDF tools are thin canvas-based hacks. This one runs PyMuPDF (the same library used by archive.org, scientific publishers, and major SaaS PDF processors) directly inside your browser via Pyodide. The Python runtime warms up in the background while you drop your file, so by the time you click an action it is ready.

Real capabilities

• Text with positions: blocks, words, full structured dict, or HTML output. Coordinates are accurate to the original PDF coordinate system.
• Real redaction: text is removed from the underlying content stream, not just covered with a rectangle. The redacted PDF cannot be un-redacted by selecting text.
• Embedded image extraction: pulls images at their original resolution and format, not page screenshots.
• Annotation listing and flattening: read every comment, highlight, and note. Flatten to bake them into the page so they cannot be edited.
• Metadata editing: set or wipe Title, Author, Subject, Keywords, Creator, Producer.
• Page operations: drag to reorder, click X to remove, rotate selected pages, extract just the pages you want, merge multiple PDFs.

Privacy

Everything runs in your browser tab. Nothing is uploaded. The Python runtime and PyMuPDF library are fetched once from a CDN and cached, so subsequent visits start instantly.