Skip to content
TakoAPI
Back to skills

vision-tagger

by sagarjhaa

DevOps & Cloud
897Downloads reported by clawskills.sh
0Views
0 likesSign in to like this skill
—Not yet rated
Your rating

Brief

Tag and annotate images using Apple Vision framework (macOS only)

Description

What This Skill Does

macOS-native image analysis using Apple's Vision framework. Detects faces, body joints, hands, text (OCR), barcodes, scene labels, and saliency regions from image files. All processing runs on-device — no cloud APIs or API keys required.

Runs entirely on-device using Apple's built-in Vision framework, so there are no API costs, no usage limits, and no image data sent to external servers.

When to Use It

  • Auto-tagging vacation photos by detected scenes and objects
  • Extracting text from scanned receipts or handwritten notes
  • Monitoring sitting posture by tracking body joint positions during work sessions
  • Scanning product barcodes from photographed inventory
  • Detecting people in security camera screenshots

Example Workflow

INPUT

User asks: Analyze this photo and tell me what objects and people are in it

AGENT
  1. 1

    Runs the compiled Swift binary against the specified image file

  2. 2

    Parses the JSON output for faces, body joints, scene labels, and text regions

  3. 3

    Maps detected labels to human-readable descriptions with confidence scores

  4. 4

    Returns a structured summary of all detected elements

OUTPUT

JSON report listing detected faces with bounding boxes and orientation angles, body skeleton joints, scene labels such as 'outdoor' and 'sky' with confidence scores, and any OCR text found in the image

Requirements

  • Xcode Command Line Tools (run: xcode-select --install)
  • Pillow Python library (run: pip3 install Pillow)

Installation

clawhub install sagarjhaa/vision-tagger

Run the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.

Links