Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/QwenLM/Qwen3-VL/llms.txt

Use this file to discover all available pages before exploring further.

​
OCR & Key Information Extraction

Qwen3-VL provides powerful optical character recognition (OCR) capabilities with expanded language support and robust performance across diverse conditions. The model excels at both general text recognition and targeted key information extraction.

​
Capabilities

​
Expanded Language Support

Qwen3-VL supports OCR in 32 languages, significantly expanded from the previous 10:
  • Latin-script languages
  • Chinese (Simplified and Traditional)
  • Japanese, Korean
  • Arabic, Hebrew, and other RTL scripts
  • Cyrillic scripts
  • Indic languages
  • And many more

​
Robust Text Recognition

The model performs reliably even in challenging conditions:
  • Low Light: Extract text from dimly lit or dark images
  • Blur: Handle motion blur and out-of-focus text
  • Tilt & Rotation: Process text at various angles
  • Rare Characters: Recognize ancient characters and specialized symbols
  • Technical Jargon: Handle domain-specific terminology and notation

​
Key Information Extraction

Beyond general OCR, Qwen3-VL can extract specific information from documents:
  • Forms: Extract field values from structured forms
  • Receipts & Invoices: Parse transaction details, dates, amounts
  • Business Cards: Extract contact information
  • IDs & Documents: Pull specific fields from identification documents
  • Tables: Extract data from tabular formats

​
Use Cases

  • Document Digitization: Convert physical documents to searchable text
  • Data Entry Automation: Eliminate manual typing from forms and receipts
  • Multilingual Content: Process documents in various languages
  • Accessibility: Make visual content accessible to screen readers
  • Search & Indexing: Enable full-text search on image-based documents

​
Try It Out

Explore OCR and key information extraction with our interactive cookbook:

OCR Cookbook

Stronger text recognition capabilities in natural scenes and multiple languages, supporting diverse key information extraction needs.
Open In Colab

​
Advanced Features

  • Natural Scene Text: Extract text from photos, signs, and real-world images
  • Handwriting Recognition: Process handwritten notes and annotations
  • Mixed Content: Handle documents with multiple languages and scripts
  • Layout Understanding: Maintain reading order in complex layouts

​
Performance Highlights

  • 32 language support (up from 10)
  • Robust in challenging conditions
  • Accurate recognition of rare and ancient characters
  • Improved long-document structure parsing

Build docs developers (and LLMs) love

Get started for freeTalk to us