Pyyan / OCR & Document AI / MinerU

MinerU

A pipeline, not a model: PDF to clean markdown.

currentOpenDataLabVerified 2026-08-13

Pipeline

PDF to markdown

Specifications

OmniDocBench~90
Open weightsYes
HandlesPDF, formulas, tables
LicenceAGPL-3.0
KindPipeline

What matters

  • Combines layout, formula and table models
  • Popular for building training corpora

Best for bulk academic PDF conversion.