A high-precision OCR tool for extracting mixed Chinese and English text from screenshots, receipts, and structured forms.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install siliconflow-image-ocr
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install siliconflow-image-ocr using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Image OCR?
The Image OCR skill is a specialized vision-language processing tool designed for AI agents within the Openclaw Skills ecosystem. By leveraging the SiliconFlow API, it enables seamless extraction of text from various visual formats including local image paths, remote URLs, and base64-encoded strings. It is particularly optimized for handling document-heavy tasks where precision in mixed-language (Chinese/English) environments is critical for success.
This skill serves as a bridge between visual data and structured text, allowing your agent to perceive and interpret information from invoices, tables, and screenshots. Unlike general-purpose chat models, this specific implementation focuses on the fidelity of text extraction, making it an essential utility for developers building automated data entry or document analysis pipelines.
Image OCR Use Cases
- Digitizing text from receipts, invoices, and financial documents for automated reporting.
- Extracting content from screenshots of technical documentation or code snippets.
- Parsing data from complex tables and forms for structured database entry.
- Reading shipping labels and identification documents in logistics workflows using Openclaw Skills.
How Image OCR Works
- The AI agent or user provides an image input via a local file path or a public URL.
- A multimodal prompt is generated to instruct the SiliconFlow vision model on the specific extraction task.
- The script securely authenticates with the SiliconFlow API using stored credentials.
- The vision-language model processes the image and identifies text blocks, including mixed Chinese and English characters.
- The skill returns a structured JSON output containing the extracted text for further downstream processing.
Image OCR Setup
To integrate this into your environment, follow these steps:
# 1. Set your SiliconFlow API Key as an environment variable
export SILICONFLOW_API_KEY="your_key_here"
# 2. Alternatively, create a secret file for persistent access
mkdir -p ~/.openclaw/secrets/
echo "your_key_here" > ~/.openclaw/secrets/siliconflow_api_key
chmod 600 ~/.openclaw/secrets/siliconflow_api_key
# 3. Verify the installation with a simple test
python scripts/paddleocr_vl.py --prompt "Extract text" --image-path /path/to/test.png
Image OCR Data Schema & Taxonomy
The skill utilizes a multimodal messaging structure to handle inputs and returns raw JSON data for maximum flexibility.
| Argument | Requirement | Description |
|---|---|---|
--prompt |
Required | Instructions for the OCR engine (e.g., "Extract all text"). |
--image-path |
Optional | Path to a local image file on the system. |
--image-url |
Optional | URL of an image hosted online. |
Output |
JSON | A structured response containing the full text extraction results. |
Image OCR Advanced Features
- Multimodal prompt support allows users to combine text instructions with visual data for more nuanced extraction.
- Dual credential management supports both environment variables and local secret files for secure deployment.
- Native support for mixed-language character sets ensures high accuracy for international business documents.
- Flexible input handling accepts local paths, URLs, and base64 streams, making it a versatile addition to your collection of Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
Circle Modular Wallets SDK Integration
A specialized toolkit for building ERC-4337 smart contract accounts with passkey authentication and gasless transaction support via Circle Modular Wallets.
HealthFit: Comprehensive Personal Health AI
A sophisticated multi-agent health management system that bridges Western medicine and Traditional Chinese Medicine to provide personalized fitness, nutrition, and wellness guidance.
Pinduoduo Hot Trend
A specialized automation skill for extracting public product data, pricing trends, and review highlights from Pinduoduo.
Douyin Short Drama Hot Trend
A professional AI skill designed to track, organize, and summarize trending Douyin short drama data and episode performance.
Use Circle Wallets
A comprehensive decision framework for selecting and implementing the optimal Circle wallet infrastructure for blockchain applications.
iQIYI
An intelligent streaming assistant for iQIYI that provides expert guidance on exclusive content, VIP features, and cross-platform viewing optimizations.
Featured*
