A sophisticated visual intelligence tool that enables AI agents to identify objects, scenes, text, and biological entities within images.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install image-recognize
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 image-recognize 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 Recognition?
The Image Recognition skill is a robust integration designed for Openclaw Skills that leverages Baidu AI's computer vision capabilities. It allows your AI agent to analyze and interpret visual content from local files, web URLs, or Base64 encoded strings.
This skill serves as a bridge between raw visual data and actionable intelligence, providing detailed classifications for a wide array of categories including landmarks, animals, plants, consumer products, and facial features. It is specifically optimized for general image classification and scene understanding, making it an essential component for agents requiring environmental awareness or visual data processing.
Image Recognition Use Cases
- Identifying specific species of plants or animals in research tasks.
- Automatically tagging and classifying product images for e-commerce workflows.
- Recognizing famous landmarks and tourist attractions from travel photos.
- Extracting contextual scene information to help AI agents understand user-uploaded media.
- Enhancing data enrichment pipelines by converting visual content into structured Markdown descriptions via Openclaw Skills.
How Image Recognition Works
- The AI agent receives an image input in the form of a file path, a remote URL, or a Base64 string.
- The system checks for the presence of the BAIDU_API_KEY in the environment or conversation context.
- The skill script processes the image, ensuring it meets size requirements (ideally under 1MB for optimal performance).
- Data is transmitted to the Baidu AI engine for deep learning-based visual analysis.
- The skill returns a structured Markdown response containing the primary identification, confidence scores, and supplementary reference information.
Image Recognition Setup
To use this skill, you must have Python 3 installed and a valid Baidu AI API Key.
# Set your API Key as an environment variable
export BAIDU_API_KEY="your-baidu-api-key-here"
# Basic usage for a local image
python3 scripts/image_recognize.py --image "/path/to/local/image.jpg"
# Usage for a network URL
python3 scripts/image_recognize.py --image "https://example.com/photo.png"
Image Recognition Data Schema & Taxonomy
The skill utilizes a simple parameter structure and returns rich Markdown data.
| Parameter | Type | Description |
|---|---|---|
| --image | String | Path, URL, or Base64 string (Required, max 4MB) |
| --similar_count | Integer | Number of similar results to return (Default: 3) |
Output Metadata Includes:
- Description: A text-based summary of the recognized subject.
- Confidence Score: A numerical rating of the recognition accuracy.
- Category Labels: Taxonomy tags such as 'Animal', 'Plant', or 'Object'.
- Reference Links: URL citations to authoritative sources like Baidu Baike.
Image Recognition Advanced Features
- Multi-format support allowing seamless transitions between local storage and web-based image processing.
- Similarity analysis to retrieve related visual matches and expanded context.
- High-confidence scoring to filter out low-probability identifications in automated Openclaw Skills workflows.
- Deep integration with search databases to provide educational or descriptive summaries of recognized items.
SKILL.md
Loading
Related Openclaw Skills
Dup File Finder
A streamlined command-line utility for identifying and managing duplicate files within any specified directory.
Math
A sophisticated mathematical engine for Openclaw Skills that adapts its pedagogical approach to match the user proficiency from elementary basics to expert research.
Multi Search Engine Integration
A high-performance search aggregation skill for AI agents that integrates 16 global and domestic search engines without requiring API keys.
Chishi Writing Style
An AI skill designed to generate pseudo-serious, absurd, and internet-native Chinese prose in the Chishi style.
Clean Filenames
A robust CLI tool designed to sanitize and normalize filenames for professional organization and cross-platform reliability.
Xiangji Translation Service
A comprehensive translation tool that leverages multiple AI engines to provide professional text and image-based language conversion.
METADATA
Featured*
