yescan-ai/yescan-skill

What does yescan-skill do?

Quark Scanner — A three-in-one document AI skill pack for OpenClaw Agent: OCR text recognition (17 scenarios), document scanning enhancement (13 modes), and image-to-Office conversion (Word/Excel/PDF). Unlock 30+ capabilities with a single API key, ready with openclaw skills install.

Overview

This skill integrates the Quark Scanning King (夸克扫描王) API into your OpenClaw Agent, giving it powerful document understanding and generation abilities. Whether you need to extract text from photos, enhance scanned documents, or convert images into editable Office files, this skill has you covered.

Features

• OCR Text Recognition: Supports 17 common document scenarios including ID cards, bank cards, invoices, receipts, handwritten notes, printed text, business cards, and more. Returns structured text with layout information.
• Document Scanning Enhancement: Apply 13 enhancement modes such as perspective correction, shadow removal, contrast adjustment, binarization, color enhancement, and noise reduction to improve document readability.
• Image-to-Office Conversion: Convert images directly into:
  • Word (.docx) – editable text documents
  • Excel (.xlsx) – spreadsheets with detected tables
  • PDF (.pdf) – searchable and compressed PDFs
• Single API Key: Access all 30+ capabilities using one API key obtained from the Quark Scanner developer portal.

Installation

openclaw skills install yescan-ai/yescan-skill

Configuration

Set the following environment variable or pass it during agent initialization:

YESCAN_API_KEY=your_api_key_here

Usage

Example 1: Recognize text from an image

result = await agent.execute("yescan-ai/yescan-skill", {
  "action": "ocr",
  "image_url": "https://example.com/receipt.jpg",
  "scenario": "receipt"
})
print(result.text)

Example 2: Enhance a scanned document

result = await agent.execute("yescan-ai/yescan-skill", {
  "action": "enhance",
  "image_url": "https://example.com/scan.jpg",
  "mode": "perspective_correction"
})
print(result.enhanced_image_url)

Example 3: Convert image to Excel

result = await agent.execute("yescan-ai/yescan-skill", {
  "action": "convert",
  "image_url": "https://example.com/table.png",
  "target_format": "xlsx"
})
print(result.download_url)

Supported Scenarios (OCR)

ID	Scenario	Description
1	id_card	Chinese ID card
2	bank_card	Bank card number and details
3	invoice	Chinese invoice (fapiao)
4	receipt	General receipt
5	handwritten	Handwritten text
6	printed	Printed book or document
7	business_card	Business card
8	passport	Passport data page
9	license_plate	License plate (Chinese)
10	vehicle_vin	Vehicle VIN number
11	table	Table with rows and columns
12	form	Structured form
13	menu	Menu or price list
14	label	Product label or barcode
15	code	QR code or barcode
16	mixed	Mixed text and image
17	general	General purpose OCR

Supported Enhancement Modes

ID	Mode	Description
1	auto	Automatic enhancement
2	perspective_correction	Correct perspective distortion
3	shadow_removal	Remove shadows
4	contrast_optimization	Optimize contrast
5	binarization	Convert to black-and-white
6	color_enhancement	Enhance colors
7	noise_reduction	Reduce noise
8	sharpening	Sharpen the image
9	deblurring	Reduce blur
10	light_correction	Correct uneven lighting
11	straighten	Straighten skewed images
12	crop	Auto crop to document borders
13	compress	Compress image while preserving quality

Limitations

• Maximum file size for images: 20 MB
• Supported image formats: PNG, JPG, JPEG, BMP, WEBP
• Images must be in base64 encoding or accessible via public URL
• OCR accuracy depends on image quality and text clarity
• Conversion to Office formats may lose complex formatting (e.g., nested tables, merged cells)

API Key

To obtain your free API key, visit the Quark Scanner Developer Portal and create an account.

License

This skill is provided under the MIT License. See LICENSE for details.