FL PDF Loader

Unremarkable by design, and that's the point: FL_PDFLoader is the plain "get one PDF file into the graph" node that everything else in Fill-Nodes' PDF group builds on. FL_PDFTextExtractor needs a loaded PDF to pull text from, FL_PDFImageExtractor needs one to pull embedded images from, FL_PDFMerger needs two of them, FL_PDFEncryptor needs one to lock - this is where that object comes from in the first place, for a single file. (If you're loading a whole directory of PDFs at once instead of one, FL_BulkPDFLoader is the sibling node for that.)

How it works

Give it a file path. It validates the file exists and is actually a valid PDF before proceeding, then reads it into a PDF data object carrying the file path, page count, and the raw binary content - everything downstream nodes need without re-reading the file from disk each time.

The inputs and outputs that matter

• pdf_path - the only input, a string path to a single .pdf file.
• Output: PDF - a Fill-Nodes-specific type, not a raw file path or bytes. It only connects to other nodes in this pack's PDF group (FL_PDFTextExtractor, FL_PDFImageExtractor, FL_PDFMerger, FL_PDFEncryptor, FL_PDFToImages, and so on) - that's deliberate, since it's carrying structured metadata (page count included) alongside the binary content, not just bytes.

That's the entire node. No configuration, no extraction logic - that all lives in whichever node you plug this into next.

How to install it

Via ComfyUI Manager: search "ComfyUI_Fill-Nodes" and install, or manually:

cd ComfyUI/custom_nodes
git clone https://github.com/filliptm/ComfyUI_Fill-Nodes

then restart ComfyUI. Expect a PyPDF2 dependency across this pack's PDF node group - this isn't pure standard-library file reading, since page-count and format validation need real PDF parsing. The pack overall is a large, general-purpose collection from filliptm, one prolific solo author, also covering image FX, GPT/Gemini/Fal API wrappers, and an entire KSampler family; the PDF tools are one self-contained slice of it.

Common issues & troubleshooting

"File not found" on a path that looks right. Custom nodes run relative to wherever ComfyUI's Python process was started, not necessarily the folder you're thinking of - use an absolute path if you're not sure.

"Invalid PDF format" on a file you're sure is a PDF. A genuinely corrupted download, a file that got renamed to .pdf without actually being one, or a PDF variant the underlying library can't parse (encrypted PDFs you haven't supplied a password for, in particular) can all trip this. Try opening the same file in a normal PDF reader first to rule out a bad file before assuming the node is at fault.

You need multiple PDFs at once, not one. This node is single-file by design - reach for FL_BulkPDFLoader if you're pointing at a whole directory instead of a single path.