Skip to main content

​
Signature

pymupdf4llm.to_json(
    doc: str | pymupdf.Document,
    **kwargs
) -> list[dict]

​
Parameters

​
doc
str | pymupdf.Document
required
Path to the document file, or an already-opened pymupdf.Document instance. Supports PDF, XPS, eBooks, and — with PyMuPDF Pro — Office formats.
​
**kwargs
various
Additional parameters are shared with to_markdown(). See the to_markdown() API reference for details.
For other parameters, see the shared to_markdown() API reference which applies to all extraction functions.

​
Returns

​
list[dict]
list
A list of page objects, one per extracted page. See JSON Schema for the full field reference.See Extract JSON for detailed block structure examples.

​
Raises

ExceptionCondition
FileNotFoundErrordoc is a path string that does not exist
ValueErrorAn index in pages is out of range for the document
ImportErrorocr=True or force_ocr=True but the ocr dependency is not installed

​
Examples

​
Minimal

import pymupdf4llm

data = pymupdf4llm.to_json("document.pdf")

​
Iterate over blocks

for page_num, page in enumerate(data.get("pages", [])):
    for block in page.get("boxes", []):
        for line in block.get("textlines", []):
            for span in line.get("spans", []):
                bbox = span.get("bbox", []) # bounding box for this text span
                text = span.get("text", "") # text content of the span
                flags = span.get("flags", 0) # font style flags (bitmask)

​
See Also

Extract JSON Guide

Full walkthrough with bounding boxes, span flags, and pipeline examples.

JSON Schema

Complete field reference for every object in the JSON output.

to_markdown()

Markdown output for LLM ingestion and readable docs.

Tables Guide

Working with table blocks in the JSON output.