Which image formats are supported by this skill?

OCR Finder supports most common image formats, including PNG and JPG, through the PIL (Pillow) library.

Why is the first run slower than subsequent uses?

The initial execution loads the EasyOCR model into memory, which typically takes 5-10 seconds. Subsequent calls use a cached reader for significantly faster performance.

Does OCR Finder require macOS accessibility permissions?

No, unlike native UI inspectors, OCR Finder processes image files directly and does not require system-level accessibility permissions.

Can I use OCR Finder to click buttons automatically?

Yes, the --click command returns the center coordinates of the detected text, which can then be piped into system automation tools to simulate mouse clicks.

How can I improve detection for low-contrast text?

You can use the --min-confidence flag to lower the threshold (e.g., --min-confidence 0.3), which helps the OCR engine capture fainter or less distinct text.

OCR Image Finder

Name: OCR Image Finder
Author: smykla-skalski

bysmykla-skalski

0•

Productivity & Workflow

Extracts text and precise click coordinates from images and screenshots using EasyOCR for automated UI workflows.

The OCR Finder skill enables Claude to interpret and interact with text within any image format, such as PNG or JPG, without requiring macOS accessibility permissions. By leveraging EasyOCR, it identifies text regions, provides detailed bounding boxes, and calculates center-point click coordinates. This tool is particularly valuable for UI automation, testing non-standard interfaces, and extracting data from visual captures where traditional UI inspection tools are restricted or unavailable.

Key Features

010 GitHub stars

02Configurable confidence thresholds to fine-tune OCR detection accuracy

03Supports substring, exact, and case-sensitive text matching filters

04Extracts text from screenshots and UI captures without accessibility permissions

05Returns structured JSON output for seamless automation pipeline integration

06Provides precise X/Y click coordinates for detected text elements

Use Cases

01Verifying visual UI elements and button placements during automated regression testing

02Automating interactions with legacy applications or custom UIs that lack accessibility labels

03Extracting text data from low-contrast or complex images for research and documentation

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add smykla-skalski/research ocr-finder

For use in Claude.ai and ChatGPT