PaddleOCR

At a glance

Input

Images and PDF

Output

JSON, Markdown

Provenance

Page and region information; pipeline-dependent

Scanned pages

Supported

Deployment

Self-hosted

Availability

Open source

Costs to plan for

Compute and model costs

Limitations to check

Basic OCR and document-structure pipelines produce different results. Select table and formula modules explicitly.

Build a meaningful sample

Include your hardest pages: low-quality scans, missing fields, dense tables, and evidence split across pages. Validate output correctness separately from citation accuracy.