Mohammad Ausaf Logo Image
Mohammad Ausaf

Searchy

AI-powered image search and management for macOS. Because photo management on Mac sucks.

Project Overview

Searchy is a lightweight image browsing and searching application built specifically for macOS. Frustrated with the limitations of existing photo management tools on Mac, I decided to create a solution that leverages AI for intelligent image search while keeping everything on-device.

Key Features

  • 🧠 AI-Powered Search: Uses CLIP model for intelligent image understanding
  • 🔒 Privacy-First: All processing happens on-device, no data leaves your Mac
  • Lightweight: Fast and efficient, designed for Mac performance
  • 🔍 Semantic Search: Find images by describing what's in them
  • 📁 Smart Browsing: Intuitive interface for exploring your photo library

Technology Stack

Languages & Frameworks

Swift (84.5%)
Python (15.3%)
CLIP Model
Core ML
SwiftUI
Machine Learning

The Problem

Mac's built-in photo management solutions often fall short for users with large image collections. Finding specific photos becomes a tedious task of scrolling through thousands of images or relying on basic metadata search. Traditional solutions are either too heavy, cloud-dependent, or lack intelligent search capabilities.

The Solution

Searchy addresses these pain points by providing:

  • Semantic Understanding: Search for "sunset over mountains" and find relevant images
  • On-Device Processing: Your photos never leave your device
  • Performance Optimized: Built with Swift for native Mac performance
  • Extensible Design: Architecture allows for future AI model integrations

Future Roadmap

  • 🗑️ Duplicate image detection and deletion
  • 🏷️ Custom AI models for specific image domains
  • 📝 User-defined image tagging system
  • 🔍 Spotlight-style search widget integration
  • 🎨 Enhanced UI with modern design patterns
  • ⚡ Performance optimizations for large libraries

Development Journey

This project represents my exploration into the intersection of AI and practical software solutions. Working with CLIP models and Core ML has been an exciting challenge, requiring me to balance machine learning capabilities with user experience design. The Swift/Python hybrid approach allows for leveraging the best of both ecosystems.