skills.rest
liushuang393
liushuang393COMMUNITY
Explore all skills by this author →
GitHub

vision

Analyze images from URLs, files, or base64 to extract descriptions, objects, and OCR text.

Updated Aug 27, 2026
Search Tags:#image#multimodal#vision#OCR#analyze#object-detection
One-click install
npx skills add https://github.com/liushuang393/serverlessAIAgents --skill vision-liushuang393
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: vision
Source: https://github.com/liushuang393/serverlessAIAgents/tree/main/agentflow/skills/builtin/vision
Command: npx skills add https://github.com/liushuang393/serverlessAIAgents --skill vision-liushuang393

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill enables automated image understanding by describing content, detecting objects, and extracting text (OCR) from images using leading vision providers.

Core Features & Use Cases

  • Image analysis: Generate natural language descriptions of image content.
  • OCR: Extract visible text from images for indexing and processing.
  • Object detection: Identify and list prominent objects within an image.
  • Multi-provider support: Works with OpenAI (GPT-4V), Claude Vision, and Gemini Vision via a unified interface.
  • Use Case: Catalog product photos by description, tags, and extracted text for accessibility and search.

Quick Start

Use the Vision Skill to analyze an image by URL or upload. Initialize VisionSkill, call analyze_image with image_url and enable extract_text and extract_objects as needed, and read the resulting description, objects, and text in the response.

Frequently Asked Questions about vision

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I extract text from an image using OCR?▼

You can extract text from an image using OCR by calling the analyze_image function with extract_text enabled. It processes URLs, local files, or base64 data and returns structured results containing the extracted text.

How does multimodal image analysis work with OpenAI, Claude, and Gemini?▼

Multimodal image analysis works by sending image inputs to vision providers like OpenAI GPT-4V, Claude Vision, or Gemini Vision through a unified interface. It returns structured results including descriptions, detected objects, and OCR text.

Can I use base64 data or local files for object detection?▼

Yes, you can use base64 data or local files for object detection. The Skill accepts local files, base64 data, and URLs as input, detecting and listing prominent objects within the provided image.

What is the best way to catalog product photos with descriptions and tags?▼

The best way to catalog product photos is using this Skill to generate natural language descriptions, extract visible text, and detect objects. It outputs structured results useful for accessibility and search indexing.

Do I need a specific vision provider configured to analyze images?▼

Yes, you need a configured vision provider to analyze images. The Skill requires httpx and a provider like OpenAI GPT-4V, Claude Vision, or Gemini Vision to process images and return structured results.

Related Skills

Explore More Design & Creative Skills→
alex-tgkalex-tgk
1

gemini-video-understanding

Analyze local and YouTube videos via the Gemini API for summaries, transcripts, and timestamped insights.

Community
Advanced
orthogonal-shorthogonal-sh
12

image-analyzer

Extract text, describe scenes, and detect objects from images.

Official
Intermediate
nexscope-ainexscope-ai
38

ecommerce.multimodal-recognize-image

Analyze image URLs with multimodal AI for visual content and OCR.

Official
Intermediate
TonyQ-AITonyQ-AI
9

mimo-image-understanding

Analyze images with Xiaomi MiMo for OCR, UI review, and frontend debugging.

Community
Advanced