Home
/
Projects
/
AI / PHOTO & VIDEO

AI VIDEO EDITOR

Modernizing an iOS Video Editor With AI Video Generation

AI VIDEO EDITOR

Modernizing an iOS Video Editor With AI Video Generation

An iOS development partnership: we stabilized a popular video editing app, launched AI generation features, built ad monetization from scratch, and rebuilt the core editor while releases kept shipping every month.

35% faster loading

for app launch and key features

Editor rebuilt in 7 weeks

without pausing releases
iOS Development
Legacy Modernization
Generative AI
Video Editing

About the product

A mobile video editor for content creators, competing in the same space as the best-known short-form editing apps. When we joined, it ranked in the top 40 of its App Store category.

Users create videos from trending templates, edit clips on a multi-track timeline with text, music, effects, and filters, and generate images and videos with AI. Premium features are offered through subscriptions and multiple paywall variants.

goals

Custom Software Development
Bring the codebase and release process back under control so the product could evolve predictably.
Custom Software Development
Refresh key screens, starting with the home screen, to match the expectations of today's creator audience.
Custom Software Development
Add AI video generation and upgrade the existing AI image generation.
Custom Software Development
Grow revenue with flexible paywalls and a new advertising channel.
Custom Software Development
Keep the app compatible with the latest iOS releases and Apple requirements.
Custom Software Development
Rebuild the core video editor for better performance and maintainability.

Needs

The app had reached the market quickly: it was built on a licensed third-party video editing engine, with templates, paywalls, and the surrounding infrastructure added on top. Over time, this layered implementation made every new feature slower and riskier to ship.

Meanwhile, competitors were moving ahead with new features and AI integrations. The product needed a stronger engineering team to stabilize the code, modernize the experience, and add the AI capabilities users now expect.
CHALLENGES

The Key Challenges We Addressed

01

Restoring Control Over the Codebase and Releases

Parallel changes lived in scattered branches with no clear path to a release. We restructured the Git flow, reviewed every pending change, and assembled a single stable version of the app. We also moved dependencies to Swift Package Manager and introduced shared code formatting. This made a steady monthly release rhythm possible.

02

Untangling Undocumented Legacy Logic

Much of the app logic had been layered over time without documentation. We traced the actual behavior, agreed on the expected one with the product team, and rewrote services step by step, starting with the app launch flow and the paywall logic.

03

Making Long-Running AI Generation Feel Smooth on Mobile

AI video generation takes 20 seconds or more, far longer than users are used to waiting in a mobile app. We designed the client flow around it: the app tracks generation status in the background, restores progress when the user returns, and sends a push notification as soon as the video is ready.

04

Modernizing Without Pausing the Product

The app had to keep shipping updates while we redesigned screens, added AI features, and rebuilt the editor. We kept a monthly release rhythm the whole time. The new editor opens existing user projects, so nothing has to be migrated. It is switched on with a remote flag, so the rollout can be gradual and reversible.
SOLUTION DEVELOPMENT

A Faster Editor, AI Generation, and a Codebase

Ready to Grow

Redesigned Home Screen and Navigation

We redesigned the home screen twice. The first iteration added a tab bar, shortcuts, templates, and video formats. The second put AI creation tools at the center, with a native Liquid Glass tab bar. We also removed a separate setup screen, so a new project now starts one step sooner.

Three AI Features

We launched AI Video, Text-to-Image on the Nano Banana model, and a fully rebuilt Text-to-Video. Users choose between Veo 3.1 and Seedance, set characters, style, sound, and camera motion, and pick a 9:16 or 16:9 format and clip length.

Generation Flow and Credits

Generation runs in the background: the app tracks its status, restores the screen when the user returns, and sends a notification when the video is ready. Every request checks subscription and credit balance, credits are returned if a generation fails, and the price updates when the user switches models.

New Video Editor Engine

In 7 weeks we rewrote the editor logic: tracks, trimming, transitions, keyframes, filters, crop, masks, speed curves, text, auto subtitles, voice recording, and export, with new undo and redo. Each tool now opens in one panel with clear tabs, and the result is visible right away in the preview.

Paywalls and Ad Monetization

Paywalls run on Superwall and can change without an app release; the native onboarding paywall follows the winning A/B variant, and subscriptions are handled by RevenueCat. We also built ad monetization from scratch: AdMob banners, GDPR consent management, and six mediation networks.

Analytics, Localization, and Platform

We moved marketing tools to Braze, cleaned up Amplitude events, and added attribution SDKs. The app now supports right-to-left languages, including Arabic. Filters, transitions, and effects download on demand, which reduces install size, and the code is adapted to iOS 27.

ACHIEVED

RESULTS

ACHIEVED

RESULTS

Core Video Editor Rebuilt in 7 Weeks

The new engine covers 20+ editing features, opens existing user projects without migration, and is enabled by a remote flag, so the switch can be gradual and rolled back at any time.

3 AI Features Launched

AI Video, Text-to-Image, and a rebuilt Text-to-Video with two models, characters, style, sound, and camera motion, all connected to subscriptions and credits.

From 0 to About 1,700 Unit Tests

The project went from no unit tests to about 1,700, mostly covering the new editor logic. Together with Swift Package Manager, code generation, and shared formatting, this gives the product a reliable base for further development.

35% Faster App and Feature Loading

A single, streamlined startup sequence, on-demand loading of filters, transitions, and effects, and a lighter asset catalog cut the time to a usable home screen by 35%. Key features, from the new editor to AI tools, open faster as well.

4x Fewer Crashes

The crash-free users rate grew as crashes dropped fourfold. Rewritten legacy services, a unified app launch flow, the new editor engine, and broad unit test coverage removed the most frequent sources of failures.
Tech stack

Tech stack

Mobile Development

Swift
UIKit
Swift Package Manager

Video Editing

Meishe SDK

Generative AI

Veo 3.1
Seedance
Nano Banana

Monetization

Superwall
RevenueCat
Google AdMob
Ad Mediation

Analytics and Marketing

Firebase
Amplitude
AppsFlyer
Braze

Configuration and Data

Firebase Remote Config
Firestore

Quality and Delivery

Swift Testing
SwiftFormat
SwiftGen
Codemagic
AI-Assisted Development
contacts@gilzor.com
Poland, Warsaw, Office 58, street Adama Mickiewicza 37, 01-625
Cyprus, Limassol, Office 113E 3030, BLOCK A, AGIOS PAVLOS, 228, Makariou III
Services
Web Development
Mobile Development
UX/UI Design
Business analysis
Mobile App Audit
AI-Built App to Production
All services →
Subscribe

By subscribing you agree to receive emails from Gilzor. You can unsubscribe at any time. Privacy Policy.

You subscribed for our blog and news. Thank you!
Oops! Something went wrong while submitting the form.
Medium
© 2026 All rights reserved
Legal information
Privacy Policy