
Retro Diffusion App
github.com/dimillian/retrodiffusionapp- Category
- AI Tools
- Rank
- No. 1552Tools index
- Pricing
- Open Source
- Type
- APP
- Builder
- dimillian
- GitHub
- 120 stars
- Date
About
Dimillian's iOS app for retrodiffusion.ai — generate pixel art and retro game assets from text prompts on your phone.
What it does
Retro Diffusion App combines two asset workflows in one native interface. It transforms existing photos into pixelated images, or creates new images from prompts across more than 30 model styles. Before submitting work, it previews the expected cost. Finished images enter a paginated local library and can also be saved to Photos.
Why it's ranked here
The focused workflow is its strongest argument. Cost previews reduce guesswork, dimensions are adjustable, and the style catalog covers spritesheets, maps, textures, portraits, interface elements, and other practical asset types. Local persistence makes results easier to revisit. However, it remains tightly coupled to one paid API and very recent Apple tooling.
What's good
Image processing and photo writes run away from the main thread. Actor isolation protects networking and local storage state, while observable clients keep interface updates straightforward. The library stores PNG files with searchable metadata in SQLite, loads records in pages, and caches decoded images. Generated work is saved automatically, reducing the chance of losing a useful result.
Tradeoffs
It requires iOS 26.1, Xcode 26.1, an API account, a key, and available credits. Missing key configuration can terminate the app during startup. Pixelation downsizes inputs to fit a 256-pixel maximum dimension and removes transparency during RGB conversion. Cancellation only removes pending work, while started requests continue. The supplied tests are mostly generated placeholders rather than behavioral coverage.
How to use it well
It suits iOS users iterating on small retro assets who want prompt generation, photo conversion, visible costs, and a local result archive together. Choose a purpose-specific style, preview the charge, then save promising outputs into the built-in library before exporting to Photos. It does not replace a pixel editor, animation tool, game engine integration, or general API client.
Technical notes+
RetroDiffusionApp/RetroDiffusionAppApp.swift injects NetworkClient, LibraryClient, and GenerationQueue into SwiftUI. RetroDiffusionApp/actors/Networking.swift sends JSON to the /v1/inferences API with an X-RD-Token header, handles cost checks, and decodes base64 images. RetroDiffusionApp/utils/ImageUtils.swift performs asynchronous resizing and RGB PNG encoding. RetroDiffusionApp/actors/LibraryStore.swift stores metadata in SQLite and PNG data on disk; RetroDiffusionApp/library/LibraryClient.swift adds paging and NSCache. RetroDiffusionApp/utils/GenerationQueue.swift tracks pending, active, completed, and failed tasks, then persists successful results. RetroDiffusionAppTests/RetroDiffusionAppTests.swift and RetroDiffusionAppUITests/RetroDiffusionAppUITests.swift contain template examples, with only launch and launch-performance behavior exercised.
Observed
- Primary language
- Swift
- Interface
- Native SwiftUI iOS application backed by the RetroDiffusion HTTP API
- Platform support
- Requires iOS 26.1 or later and Xcode 26.1 or later
- Installation surface
- Source project configured and built through Xcode; no packaged store distribution is described
- Authentication
- RetroDiffusion API key loaded from user preferences or a gitignored property list
- Persistence
- SQLite metadata store with PNG files on disk, pagination, and an in-memory image cache
- Test structure
- Separate unit and UI test targets exist, but their supplied cases are mostly template placeholders
Read from README.md, RetroDiffusionApp/RetroDiffusionAppApp.swift, RetroDiffusionAppTests/RetroDiffusionAppTests.swift, RetroDiffusionAppUITests/RetroDiffusionAppUITests.swift, RetroDiffusionAppUITests/RetroDiffusionAppUITestsLaunchTests.swift, RetroDiffusionApp/models/Models.swift, RetroDiffusionApp/utils/Constants.swift, RetroDiffusionApp/utils/ImageSaver.swift, RetroDiffusionApp/utils/ImageUtils.swift, RetroDiffusionApp/actors/Networking.swift, RetroDiffusionApp/actors/LibraryStore.swift, RetroDiffusionApp/library/LibraryClient.swift, RetroDiffusionApp/models/GenerationTask.swift, RetroDiffusionApp/utils/GenerationQueue.swift, RetroDiffusionApp/networking/NetworkClient.swift.
What it can do
Generate pixel art from text prompts
Text description → Pixel art image
Create retro game sprites from text descriptions
Text prompt describing game character or object → Retro-style sprite image
Generate retro game backgrounds from text
Text description of scene or environment → Pixelated background image
Create game tiles from text prompts
Text description of terrain or objects → Tileable pixel art assets
Generate retro-style UI elements from descriptions
Text prompt describing interface elements → Pixel art UI components
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.