Beyond the Lens.
An open source AI-native camera stack for iOS.
Website · iOS Client · Cloudflare Server · Quick Start · Contribute
YOYO Camera is a full-stack reference implementation for the next generation of mobile photography: pro camera controls, real-time film rendering, context-aware automation, and generative AI workflows in one native iOS app.
If you are building camera apps, AI creative tools, mobile rendering pipelines, or Cloudflare-backed consumer products, this repo is designed to be studied, forked, and improved.
- Real app architecture, not a toy demo: native iOS client, edge backend, auth, subscriptions, credits, tests, localization, and production-style modules.
- AI before and after capture: scene-aware inspiration while shooting, plus AI darkroom workflows after capture.
- Automation-first camera UX: build rules that react to scene, lighting, composition, color, motion, time, location, and capture events.
- Film rendering with character: LUTs, Core Image, Metal shaders, grain, halation, bloom, vignette, and film-inspired presets.
- Useful for makers: a practical codebase for learning AVFoundation, SwiftUI camera UX, Core ML/Vision analysis, and Cloudflare Workers APIs.
| Area | What YOYO Includes |
|---|---|
| Camera | Manual exposure, focus, ISO, shutter, white balance, zoom, photo, video, Live Photo, histogram, and quick capture |
| Intelligence | Scene classification, object detection, composition scoring, lighting analysis, color analysis, and motion-aware rules |
| Automation | Rule engine for context-aware camera behavior |
| Rendering | Built-in filter recipes, film presets, LUT assets, custom filters, and Metal-powered film effects |
| AI | Inspiration generation, AI image generation, object removal, portrait enhancement, blur repair, ID photo, color grading, and closed-eye fixes |
| Backend | Apple Sign In, JWT auth, subscriptions, credits, automation share codes, Google GenAI integration, D1 schema, and App Store notifications |
YOYO explores a simple question:
What if the camera understood what you are trying to shoot and helped you make better creative decisions in real time?
That idea shows up in four product loops:
- Shoot with intent: pro controls, gesture-friendly camera UI, quick actions, widgets, and App Intents.
- Understand the scene: analyze composition, light, color, objects, histograms, and context on device.
- Automate the setup: trigger camera actions from rules instead of repeatedly adjusting settings by hand.
- Create with AI: get visual inspiration from the current scene, then refine images with AI darkroom operations.
- SwiftUI camera interface built on AVFoundation capture sessions.
- Manual exposure, shutter speed, ISO, focus, white balance, lens, zoom, and audio/video controls.
- Real-time preview pipeline with live filters, histogram support, timers, guidelines, and orientation handling.
- Photo, video, and Live Photo capture with metadata and save services.
- Quick launch through home screen shortcuts, App Intents, and WidgetKit.
- "If this, then that" rule system for camera behavior.
- Conditions for scene type, lighting, composition, subject, object count, motion, histogram state, color, time, date, location, altitude, and capture events.
- Actions for camera settings, filter changes, exposure behavior, focus behavior, and UI feedback.
- Shareable automation rules backed by short codes from the server.
- Data-driven built-in filter registry with dozens of named looks.
- Film preset model inspired by color negative, cinema, portrait, landscape, and instant film characteristics.
- Core Image and Metal rendering for cine tone, grain, halation, bloom, fog, vignette, channel mixing, and highlight rolloff.
- LUT-backed assets and custom filter management for extending the visual system.
- AI Inspiration: upload a scene preview and receive creative directions, generated images, and style prompts.
- AI Darkroom: process captured photos with operations such as object removal, portrait enhancement, professional photo, social avatar, blur repair, color grading, and eye fixes.
- Credit-aware UX: backend-managed credits and subscription status for paid AI features.
- Bilingual product foundation: English and Simplified Chinese localization assets.
| Client | Server |
|---|---|
| SwiftUI | TypeScript |
| AVFoundation | Cloudflare Workers |
| Core Image | Cloudflare D1 |
| Metal | Wrangler |
| Core ML and Vision | Google GenAI |
| StoreKit | JWT auth |
| App Intents and WidgetKit | Vitest |
| Firebase Analytics | App Store Server API |
git clone https://github.com/Day1Labs/YOYO-Camera.git
cd YOYO-CameraRequirements:
- Xcode 15 or later
- iOS 17 or later
- Swift 5.9 or later
- A valid Apple developer setup for device testing
- A Firebase
GoogleService-Info.plistif you enable analytics/auth-dependent flows
Open the project:
open YOYO-Client/Yoyo.xcodeprojThen select the Yoyo scheme, configure signing, and build on a real device or simulator.
Optional CLI build:
cd YOYO-Client
xcodebuild -project Yoyo.xcodeproj -scheme Yoyo buildRequirements:
- Node.js 18 or later
- Cloudflare account
- Wrangler access
Install and start:
cd YOYO-Server
npm install
npm run devRun tests:
npm run testDeploy:
npm run deployApply the local D1 schema when needed:
wrangler d1 execute DB --file=schema.sql --localThe server uses Wrangler vars/secrets for auth, billing, and AI features. Configure values such as:
JWT_SECRETGEMINI_API_KEYIAP_SHARED_SECRETAPP_STORE_PRIVATE_KEYAPP_STORE_BUNDLE_IDAPP_STORE_ISSUER_IDAPP_STORE_KEY_ID
The app currently points AI services at https://yoyo.day1-labs.com. Update client service base URLs if you run your own backend.
- Building a modern mobile camera experience with SwiftUI and AVFoundation.
- Designing real-time camera controls that feel native and responsive.
- Creating Core Image and Metal pipelines for live visual effects.
- Combining Core ML/Vision analysis with rule-based product automation.
- Structuring AI-powered mobile features across app and backend.
- Implementing Apple Sign In, subscriptions, credits, and feature gating.
- Shipping a Cloudflare-native backend for a consumer mobile product.
Issues, ideas, and pull requests are welcome. See CONTRIBUTING.md for development setup, pull request checklist, localization notes, and good first contribution areas.
- Website: https://day1-labs.com/yoyo/
- iOS client:
YOYO-Client - Cloudflare backend:
YOYO-Server - Contributing guide:
CONTRIBUTING.md - License:
MIT
Released under the MIT License. See LICENSE for details.
If YOYO Camera helps you learn, prototype, or ship faster, a star is appreciated and helps more makers discover the project.