ComfyUI PyMuPDF
ComfyUI nodes for loading PDFs and extracting images/text with PyMuPDF.
Nodes (7)
Images, text, and metadata in one pass
Feed your document to an LLM without the pixels
The metadata and table of contents reader
Read the hyperlinks and sticky notes
The front door to every other node in this pack
Pages as an IMAGE batch, ready for img2img
Get the extracted text out of the graph and onto disk
ComfyUI PyMuPDF
ComfyUI custom nodes for loading PDFs and extracting PDF images, rendered pages, text, metadata, links, and annotations with PyMuPDF.
Nodes
Load PDF (PyMuPDF): upload or select one PDF. Batch mode returns all PDFs inComfyUI/input/pymupdf.Extract PDF (PyMuPDF): extract embedded images and text with page/image limits for large documents.Extract PDF Text (PyMuPDF): extract text, blocks, words, HTML, XHTML, XML, or JSON without rendering images.Render PDF Pages (PyMuPDF): render selected pages to a ComfyUIIMAGEbatch.PDF Info (PyMuPDF): read PDF metadata and table of contents.PDF Links & Annotations (PyMuPDF): export page links and annotations as JSON.Save PDF Text (PyMuPDF): save extracted text or JSON to the ComfyUI output directory.
Upload UX
The loader node supports a PDF upload button, dropping PDF files onto the node, and pasting PDF files while the node is selected. It also supports local PDF paths copied to the clipboard.
Languages
The plugin includes ComfyUI locale files for Chinese node names, inputs, outputs, and descriptions. When ComfyUI is set to English, nodes use English names. When the global language is Chinese, ComfyUI loads the Chinese translations.
Install
Install from ComfyUI-Manager after the node is published to Comfy Registry, or clone this repository into ComfyUI/custom_nodes.
Install requirements in the same Python environment used by ComfyUI:
pip install -r requirements.txt
Restart ComfyUI after installation.