Tidy is Back: Our Complete Rebuild Story

v2.0 NEW RELEASE

If you are reading this, there is a chance you remember Tidy. Maybe you used it back in 2018 or 2019. Maybe you swiped through hundreds of photos, cleaning your gallery with a flick of your thumb. Maybe you left us a review. If you did, thank you. We read every single one.

And then we disappeared.

For six years, Tidy sat untouched in the App Store. No updates. No new features. No communication. The app still worked, mostly, but the world around it kept moving. iOS evolved. Phones got better cameras (and bigger photo libraries). New apps came and went. And Tidy just… sat there.

Today, that changes. Tidy v2.0 is a complete rebuild from scratch. Not an update. Not a refresh. An entirely new app, built from the ground up with modern technology, new features, and everything we have learned in the years since we first launched.

This is the story of how we got here.

The Original Tidy: 4.8 Stars and 723 Reviews

Tidy launched in 2018 with a simple premise: cleaning your photo library should feel like a game. Swipe right to keep a photo. Swipe left to drop it. No complicated menus, no folder systems, no organizational frameworks. Just quick, instinctive decisions that clear the clutter.

The response surprised us. Within months, Tidy had accumulated hundreds of reviews with an average rating of 4.8 out of 5 stars. By the time we reached 723 reviews, the pattern was clear: people loved the swipe mechanic. They loved how fast it was. They loved that it made a boring task feel fun.

Tidy by the Numbers

4.8 Star Rating

723 Reviews

31 Languages

2018 Launch Year

Users told us things like “this is the app I did not know I needed” and “I cleared 3,000 photos in one afternoon.” The app was translated into 31 languages. It appeared in App Store features in multiple countries. For a small indie app with a focused purpose, it punched well above its weight.

The original Tidy was built in Swift 4, using the MVC architecture pattern that was standard for iOS development at the time. It was a solid app for its era. But technology moves fast, and by 2019, the cracks were starting to show.

Why We Went Silent for 6 Years

The honest answer is: life happened.

Tidy was built by a small team, and when priorities shifted - as they do for indie developers - the app fell to the bottom of the list. There was no dramatic story, no acquisition, no shutdown decision. Just the slow drift that happens when a side project stops getting attention.

Meanwhile, the gap between where Tidy was and where it needed to be kept growing:

SwiftUI arrived in 2019 and matured rapidly, making UIKit-based interfaces feel dated
Photo libraries exploded in size. The average iPhone user went from 2,000-5,000 photos to 10,000-50,000
AI and machine learning went from experimental to expected. Users started expecting apps to be smart
iOS widgets and Live Activities created new surfaces that apps were expected to support
Privacy expectations increased dramatically, with users demanding on-device processing and transparency
Swift itself evolved through versions 5.0 through 6.0, with strict concurrency and new language features

By the time we seriously looked at Tidy again, the original codebase was archaeological. Swift 4 code that would not compile with modern toolchains. UIKit patterns that had been superseded twice over. An architecture that could not support any of the features users were now expecting.

We had two choices: try to patch and update the old codebase, or start fresh.

The Decision to Rebuild From Scratch

We chose to rebuild. Here is why.

Updating the old Tidy would have meant porting Swift 4 code to Swift 6, rewriting the entire UI from UIKit to SwiftUI, replacing the MVC architecture with something modern, adding new Apple frameworks (Vision, Core ML, StoreKit 2, WidgetKit), and somehow making all of these changes work together without breaking the existing app.

In software development, there is a famous essay by Joel Spolsky called “Things You Should Never Do, Part I,” which argues that rewriting from scratch is almost always a mistake. And for most projects, he is right. But Tidy was the exception:

Zero shared code - After evaluating the old codebase, we estimated that less than 5% of the original code could be reused. The language had changed too much, the frameworks were completely different, and the architecture was incompatible
Different platform - The original Tidy targeted iOS 11. The new Tidy targets iOS 18. These are effectively different platforms with different capabilities, design languages, and user expectations
Expanded scope - The original Tidy did one thing: swipe to clean. The new Tidy needed AI filters, a photo editor, compare mode, gamification, widgets, Live Activities, and more. Trying to bolt all of that onto the old architecture would be like adding floors to a one-story house by stacking them on the roof
Clean foundation for the future - Starting fresh meant we could adopt Swift 6 strict concurrency from day one, use SwiftData instead of Core Data, implement proper Clean Architecture, and build a foundation that would support years of future development

The rebuild was not a quick decision. We spent weeks evaluating the old code, prototyping in both directions (update vs. rewrite), and ultimately concluded that a ground-up rebuild would be faster, cleaner, and result in a better app than any attempt to modernize the original.

Ready to clean your gallery?

Download Tidy and start swiping. AI-powered duplicate detection, smart filters, and a photo editor built in.

App Store Google Play

Everything New in Tidy v2.0

So what did we build? Here is every major feature and improvement in the new Tidy.

SwiftUI Rewrite for Modern iOS

The entire app is built with SwiftUI, Apple’s modern declarative UI framework. This means fluid animations, native support for Dynamic Type (accessibility text sizing), automatic Dark Mode support, and a consistent, polished look across every screen. The UI feels like it belongs on iOS 18 because it was built specifically for it.

AI-Powered Smart Filters

This is the biggest addition to Tidy’s core experience. Instead of manually scrolling through your entire photo library, you can now use AI-powered filters to surface specific types of photos that are likely candidates for cleanup:

Blurry Photos - AI identifies photos with motion blur, camera shake, or missed focus
Low Quality - Detects poorly exposed, over-compressed, or otherwise low-quality images
Similar Photos - Groups nearly identical photos together using perceptual hashing, so you can pick the best and drop the rest
Duplicates - Finds exact and near-exact duplicates based on metadata and visual similarity
Best Shots - In a group of similar photos, AI suggests the sharpest, best-composed option
Utility - Identifies screenshots, receipts, documents, and other “utility” photos that you probably do not need to keep forever

All AI processing happens on your device using Apple’s Core ML and Vision frameworks. No photos are uploaded to any server. The analysis runs in the background while your phone is charging overnight, so it is ready when you are.

Full Photo Editor With 29 Filters

Tidy v2.0 includes a complete photo editing suite. This is not a basic brightness/contrast editor - it is a five-category professional editing system:

Adjust - 10 adjustment sliders including exposure, highlights, shadows, contrast, saturation, warmth, sharpness, grain, and background blur
Beauty - 6 natural beauty tools with body-positive messaging (Skin Smooth, Portrait Light, Eye Brighten, Face Slim, Eye Enlarge, Lip Plump)
Retouch - Healing brush and teeth whitening tools for targeted corrections
FX - 29 creative filters ranging from subtle film emulations to bold artistic effects
AI - Super-resolution upscaling and AI-powered style transfer effects

The editor includes undo/redo, a before/after comparison slider, and pinch-to-zoom on the canvas. Every edit is non-destructive, and you can save as a new copy or replace the original.

Compare Mode

One of the most requested features for the original Tidy was the ability to compare similar photos side by side. Tidy v2.0 delivers this with a dedicated Compare Mode that includes a thumbnail strip for quick browsing, pinch-to-zoom for detail inspection, and a before/after slider for evaluating edits.

Gamification

The original Tidy made cleaning feel fun through the swipe mechanic. Tidy v2.0 takes that further with a full gamification system:

Session stats - Photos reviewed, kept, dropped, and time for every session
Cleaning streaks - Daily and weekly streak tracking to build momentum
Milestones - Achievement system with confetti animations and haptic celebrations
Shareable cleanup cards - Beautiful, ready-to-share summary graphics for social media

20 Languages Including RTL

Tidy v2.0 launches with support for 20 languages: English, Hebrew, Arabic, Persian, Russian, French, Hindi, Romanian, Spanish, Portuguese, German, Italian, Japanese, Korean, Dutch, Polish, Thai, Turkish, Chinese Simplified, and Chinese Traditional.

Three of these languages (Hebrew, Arabic, and Persian) are right-to-left (RTL). Every screen in Tidy is designed to flip correctly for RTL users, including the swipe game, the editor, the before/after slider, and all navigation elements. RTL support is not an afterthought - it is baked into the foundation.

Interactive Widgets and Live Activities

Interactive Widgets let you start a cleaning session from your Home Screen with a single tap. Live Activities show your cleanup progress on the Lock Screen and Dynamic Island in real time. Together, they reduce the friction between “I should clean my photos” and actually doing it.

Gold Subscription

Tidy v2.0 is free to download and use. The core swipe game, basic filters, and photo comparison are all available at no cost. For users who want more, Tidy Gold unlocks:

All AI-powered smart filters (blurry detection, similar photos, best shots, utility)
Full photo editor with all 29 filters and beauty tools
Advanced gamification features
No advertisements

Gold is available as a 1-month, 3-month, or 6-month subscription. We chose this model because it aligns our incentives with yours: we only succeed if you keep finding Tidy valuable enough to keep using.

Technical Challenges of the Rebuild

Rebuilding an app from scratch sounds liberating, and it is - at first. But it also comes with its own set of challenges. Here are some of the most significant technical hurdles we faced.

Swift 6 Strict Concurrency

Swift 6 introduced strict concurrency checking, which means the compiler enforces thread safety at build time. This is great for preventing bugs but incredibly demanding during development. Every piece of shared mutable state needs to be protected by an actor, marked as Sendable, or isolated to the main thread.

For a photo app, this is especially challenging because photo processing is inherently concurrent. Loading thumbnails, running AI analysis, applying filters, saving edits - all of these operations happen on background threads while the UI updates on the main thread. Getting the concurrency model right took significant effort, but the result is an app that is virtually crash-free when it comes to threading issues.

Perceptual Hashing for Similar Photo Detection

Finding similar photos sounds straightforward until you try to implement it at scale. Tidy uses a technique called perceptual hashing (pHash) to compare photos visually. Each photo is reduced to a 64-bit hash that represents its visual content, and photos with similar hashes are grouped together.

The challenge was tuning the similarity threshold. Too strict, and related photos are not grouped together. Too loose, and unrelated photos end up in the same group. We settled on a Hamming distance threshold of 10 for standard comparisons and 14 for burst photos (taken within 10 seconds of each other), with a maximum group diameter of 14 to prevent clusters from growing too large.

We also had to choose the right clustering algorithm. A simple greedy approach would sometimes split related photos into separate groups. We implemented Union-Find clustering instead, which correctly identifies connected components and only applies the diameter constraint on oversized groups. The result is natural, intuitive photo groupings.

On-Device AI Processing

Running AI models on-device presents a constant tradeoff between accuracy and performance. A model that takes 2 seconds per photo is unusable when you have 30,000 photos to analyze. A model that runs in 10 milliseconds per photo might not be accurate enough to be useful.

We solved this with background processing. Tidy uses iOS BGProcessingTasks to run the full AI analysis overnight while your phone is charging and connected to WiFi. The analysis includes blur detection, quality scoring, similarity hashing, and face detection - all processed at high accuracy because there is no time pressure. By morning, your results are ready.

For real-time features (like the beauty editor), we use optimized pipelines that sacrifice some quality for speed. The beauty tools run at interactive frame rates because they use Apple’s Metal GPU shaders and Core Image filters, which are hardware-accelerated on all modern iPhones.

Metal Shaders for Face Warping

The beauty tools that reshape facial features (Face Slim, Eye Enlarge, Lip Plump) use Metal CIWarpKernel shaders for real-time mesh warping. Writing these shaders required working in a constrained subset of the Metal Shading Language that is compatible with Core Image’s pipeline.

One particularly tricky bug involved the jawline warping. The initial implementation positioned the warp anchor at the wrong vertical coordinate (the forehead instead of the chin), causing Face Slim to reshape the wrong part of the face. It took careful study of Vision framework’s face landmark coordinate system to get the positioning right.

RTL Layout Support

Supporting right-to-left languages is about more than flipping text alignment. Every interactive element needs to work correctly in both directions. The swipe game (“swipe right to keep, left to drop”) keeps its semantic meaning in RTL - the gestures stay the same because they are based on keep/drop semantics, not left/right direction. The before/after slider in the photo editor correctly flips its orientation. Navigation flows reverse. And every layout constraint was tested in both LTR and RTL modes.

What Is Coming Next

Tidy v2.0 is a foundation, not a finish line. Here is a preview of what we are working on for upcoming releases:

Crop and rotate tools - Coming in v2.1, with straighten guides and aspect ratio presets
Siri and App Intents integration - “Hey Siri, start a photo cleanup session” will work directly
Video compression - Reduce the size of large video files while maintaining quality
Cleanup challenges - Weekly and monthly challenges with themed goals (e.g., “Screenshot Purge Week”)
Space forecast - AI predictions for when your phone will run out of storage based on your photo-taking patterns
More AI filters - Including a meme detection filter and improved categorization

We are also exploring integration with Apple’s Foundation Models API for on-device natural language search (e.g., “show me photos of my dog at the beach”), though this depends on Apple’s rollout timeline for the API.

Our goal is to release meaningful updates every 4-6 weeks. We are not going silent again.

Thank You to Our Community

Before we wrap up, we want to say something directly to the people who used the original Tidy and might be reading this.

Thank you for the reviews. Thank you for the feature requests. Thank you for the bug reports. Thank you for the tweets and the messages and the recommendations to friends. You built something special around a simple app, and we are grateful for it.

We know that six years of silence is a long time. We know that trust, once broken, takes effort to rebuild. We are committed to earning it back - not with promises, but with consistent updates, transparent communication, and an app that keeps getting better.

If you used the original Tidy, we would love to have you back. Download v2.0, give it a try, and let us know what you think. Your feedback shaped the first version of Tidy, and it will shape every version that follows.

And if you are new to Tidy - welcome. You are joining at the beginning of something we are really proud of. We hope you like it.