Qimage-analyzer

Purpose

Uses Claude's built-in vision capabilities to analyze images. Performs a wide range of visual analysis including screenshot description, text extraction (OCR), diagram interpretation, chart analysis, UI/UX feedback, and document parsing.

Supported Analysis Types

Type	Description	Example Use Cases
OCR	Extract text from images	Document scans, screenshot text
UI Analysis	Evaluate screen structure and UX	App screenshots, web pages
Diagram Interpretation	Flowcharts, ERDs, architecture	Design documents, whiteboard photos
Chart/Graph Analysis	Interpret data visualizations	Reports, dashboards
Wireframe Analysis	Understand UI design structure	Planning documents, Figma screenshots
Document Analysis	Parse tables, forms, reports	PDF screenshots, spreadsheet captures
Comparison Analysis	Compare differences between two images	Before/after design changes, A/B tests

Workflow

Step 1: Receive Image

Ways to receive an image from the user:

• Image in Vault: Share as ![[image-name.png]] format
• Direct attachment: Drag and drop image into chat
• Provide path: Absolute or relative file path

Step 2: Identify Analysis Purpose

Determine the analysis purpose from the user's request:

• Simple description → Describe overall structure and content
• Text extraction → Extract text only in OCR mode
• UI/UX feedback → Suggest improvement points
• Data analysis → Interpret chart/table figures
• Structure identification → Describe diagram/flow

Step 3: Read Image

Read image file using the Read tool:
- Supports PNG, JPG, JPEG, GIF, WebP
- File path: use relative path within the Vault

Step 4: Perform Analysis and Provide Results

Description analysis output format:

## Image Analysis Results

### Overview
[Overall description of the image]

### Key Components
1. [Component 1]
2. [Component 2]

### Details
[Detailed analysis]

### Insights / Suggestions
[Improvement points or notable observations]

OCR output format:

## Extracted Text

[Raw extracted text as-is]

---
*Extraction confidence: [High/Medium/Low]*
*Uncertain characters: [noted]*

UI/UX analysis output format:

## UI/UX Analysis

### Screen Structure
- Layout: [description]
- Key components: [list]

### Strengths
- [Strength 1]
- [Strength 2]

### Improvement Suggestions
- [Suggestion 1]
- [Suggestion 2]

### Accessibility Issues
- [Issue 1]

Usage Examples

# Basic analysis
User: Describe this screenshot
User: What is in this image?

# OCR
User: Extract text from this document
User: Read the contents of this receipt

# Diagram
User: Explain this architecture diagram
User: How does this flowchart work?

# UI analysis
User: Give me UX feedback on this app screen
User: Identify the structure of this wireframe

# Chart analysis
User: Explain the trend in this graph
User: Summarize the data in this table

# Comparison
User: Find the differences between these two images

Limitations

• OCR accuracy may be low for very small text or blurry images
• Handwriting recognition accuracy may vary
• Complex formulas or special symbols may be inaccurate
• Processing time increases with larger image sizes

Vault Image Paths

Images in the Obsidian Vault are typically found at:

• triphos/*/images/
• images/ or attachments/ inside each project folder