skills.rest
AIAsys
AIAsysOFFICIAL
Explore all skills by this author →
GitHub

PaddleOCR 文档提取

Convert scanned PDFs and images to Markdown while preserving layout, tables, and image positions.

21|12|Updated May 17, 2026
Search Tags:#ocr#pdf extraction#markdown conversion#paddleocr#knowledge ingestion#document layout#scanned document
One-click install
npx skills add https://github.com/AIAsys/AIASys --skill paddleocr-aiasys
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: PaddleOCR 文档提取
Source: https://github.com/AIAsys/AIASys/tree/main/apps/backend/capability_sources/builtin/skill/paddleocr-skill
Command: npx skills add https://github.com/AIAsys/AIASys --skill paddleocr-aiasys

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires requests, and includes scripts (resource) components.

What problem does it solve?

This Skill solves the issue of losing critical document structure, including tables, layout, and embedded images, when extracting text from scanned PDFs or image-based documents, which makes the content difficult to edit, reuse, or ingest into knowledge systems.

Core Features & Use Cases

  • Layout-Preserving Conversion: Converts PDFs and common image formats (PNG, JPG, JPEG, BMP, TIFF) to structured Markdown while retaining original document layout, table structures, and image positions.
  • Document Correction: Automatically corrects tilted, inverted, or warped scanned documents (such as curved book pages) via optional flags to improve extraction accuracy.
  • Fallback Support: Automatically falls back to lighter OCR tools if no PaddleOCR API token is configured, ensuring the workflow does not break for users without API access.
  • Use Case: Researchers processing scanned academic papers can use this Skill to retain original table structures and image positions, so the extracted content can be directly ingested into knowledge bases or used for further analysis.

Quick Start

Use the PaddleOCR 文档提取 skill to convert the uploaded scanned document 'thesis_chapter1.pdf' into structured Markdown while preserving all tables and images.

Frequently Asked Questions about PaddleOCR 文档提取

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I extract text from scanned PDFs while preserving the original layout and tables?▼

To extract text from scanned PDFs while preserving layout and tables, you can convert the document into structured Markdown. This process retains original table structures, formatting, and embedded image positions for easy editing or knowledge base ingestion.

Can I convert image formats like PNG or TIFF to structured Markdown?▼

Yes, you can convert image formats like PNG, JPG, JPEG, BMP, and TIFF to structured Markdown. The conversion captures the original document layout and chart recognition outputs, making the text reusable for academic or archival purposes.

Does the OCR conversion work on tilted or warped scanned documents?▼

OCR conversion does work on tilted or warped scanned documents by applying automatic orientation and warping correction. This corrects curved book pages and inverted scans to ensure accurate text extraction from physical document copies.

Do I need a PaddleOCR API token to perform document layout parsing?▼

You do not strictly need a PaddleOCR API token to perform document layout parsing. The workflow automatically falls back to lighter OCR tools if no API token is configured, ensuring the document extraction process continues without interruption.

What is the best way to ingest scanned academic papers into a knowledge base?▼

The best way to ingest scanned academic papers into a knowledge base is converting them to structured Markdown. This approach digitizes the content while keeping original tables and image positions intact for downstream analysis.

Why does my scanned PDF extraction lose table structures and image positions?▼

Scanned PDF extraction loses table structures and image positions when using basic text extraction methods. Applying layout-preserving document parsing with chart recognition ensures the original document structure remains intact during the conversion process.

Related Skills

Explore More Education & Research Skills→
xongtaoxongtao

image-ocr

Extract text from PNG, JPEG, TIFF, and BMP images via Tesseract OCR.

Community
Intermediate
Arry8Arry8
1

rag-store

Ingest documents into a local ChromaDB vector store and query semantically.

Community
Advanced
fikriaffikriaf

ocr-and-documents

Extract text from PDFs and scanned documents using OCR and parsing tools.

Community
Intermediate
anilcan-karaanilcan-kara

ocr-and-documents

Extract text from PDFs and scanned documents using pymupdf or marker-pdf.

Community
Intermediate