OCR Text Recognition
Upload an image or PDF to extract text with AI-powered OCR. Supports 111+ languages with structured Markdown output.
Drag and drop your file here, or click to select
Supports JPEG, PNG, GIF, WebP, BMP, TIFF, HEIC, PDF
Maximum file size: 20MB
Batch upload: up to 20 files
Paste an image or image URL from clipboard
How It Works
Our OCR tool uses PaddleOCR-VL to analyze your images and PDFs, detecting text regions and extracting content with high accuracy. Supports 111+ languages with structured Markdown output.
How to Use
- 1
Upload your file
Drag and drop an image or PDF into the upload area, click to select a file from your device, or paste an image URL. Supports JPEG, PNG, GIF, WebP, BMP, TIFF, and PDF up to 20MB.
- 2
Choose recognition mode
Select Simple OCR for plain text extraction (1 credit/page), or Formatted OCR for structured Markdown output with tables, headings, and formulas preserved (10 credits/page).
- 3
Start recognition
Click start and our AI will process your file. Small images complete in seconds; large PDFs are processed asynchronously with status updates.
- 4
Review and export
View the extracted text in Markdown format. Copy to clipboard, or download the results for further use. Your OCR history keeps results accessible for later review.
Supported Input Formats
- JPEG (.jpg, .jpeg)
- PNG (.png)
- GIF (.gif)
- WebP (.webp)
- BMP (.bmp)
- TIFF (.tif, .tiff)
- PDF (.pdf) — up to 20MB
Frequently Asked Questions
- What languages does the OCR support?
- Our OCR tool supports 111+ languages including English, Chinese, Japanese, Korean, Arabic, Russian, and most European and Asian languages. The system automatically detects the language in your document.
- What is the maximum file size?
- The maximum file size is 20MB. Supported formats include JPEG, PNG, GIF, WebP, BMP, TIFF, and PDF.
- What is the difference between Simple and Formatted mode?
- Simple mode extracts plain text without formatting. Formatted mode preserves document structure including headings, tables, lists, and code blocks in Markdown format.
- How accurate is the OCR?
- Accuracy depends on image quality, text clarity, and language. Clear, well-lit images with standard fonts typically achieve 95%+ accuracy. Handwritten text and low-resolution images may have lower accuracy.
- Is my data kept private?
- Files uploaded via the web tool are processed and then deleted. API users' files are treated as transient — held only long enough to process the request. See our privacy controls for data deletion options.
- How many credits does OCR cost?
- Each OCR request costs credits based on the mode and number of pages. Formatted mode costs more than Simple mode. Signed-in users with credits get priority access; guest users can use the tool with a daily limit.