alibaba/Logics-Parsing

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1402 stars · 121 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Logics-Parsing is an end-to-end document parsing model from Alibaba that converts document images into structured output using a single multimodal LLM, handling complex layouts, scientific formulas, chemical structures (SMILES), tables, and handwritten text. Version 2 extends this to Parsing-2.0 scenarios such as musical sheets, flowcharts, and code/pseudocode blocks.

Use cases

• parse scanned pdf documents into structured markdown
• extract tables and formulas from document images
• convert chemical structure diagrams to SMILES
• ocr handwritten documents
• parse flowcharts and code blocks from images
• build rag pipelines over document corpora
• benchmark document parsing models on OmniDocBench

When to choose

• you need end-to-end document image to structured text without multi-stage pipelines
• documents contain complex layouts, formulas, or chemical structures
• you want a single-model deployment for document parsing

When to avoid

• you need lightweight CPU-only OCR without GPU inference
• you only need plain text extraction from digital PDFs
• you need a mature commercial document service with support

Facets

library · maturity active

ocr nlp machine-learning pdf parser computer-vision pdf artificial-intelligence deep-learning python document-parsing document-understanding multimodal-llm layout-analysis table-recognition formula-recognition smiles end-to-end natural-language-processing gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="alibaba/Logics-Parsing")