ARKit & AI Crash Course: Build iOS 18 Apps with SwiftUI

Start a new Xcode 16 project named HandGasherDetector, set the organization to example.com, enable git, and prepare the folder structure with no tests or data storage for future ARKit development.

Configure an ARKit‑only app by lowering the deployment target to iOS 16 and adding privacy camera usage and ARKit device capabilities; test on a device with a hello world example.

Explore implementing a session handler for ARKit in SwiftUI, resolve safe area issues, and render two finger tip views that track the thumb and index finger in ARViewContainer.

Introduce a session handler class that conforms to NSObject for the AR session delegate, as an observable object injected into the view, creating two green finger dots (thumb and index).

Rename methods to accurately reflect their actions, implement a create circle for fingers view inside the ARV container, and center the green indicators by positioning them at mid screen coordinates.

Explore the vision framework session handling for iOS AR with a hand pose request, process camera frames in background, extract thumb tip and index tip, and render green fingertip markers.

Detect pinch gestures by the distance between the thumb tip and index tip, then attach a 10 centimeter red mesh to a horizontal plane using ARKit raycasting.

Fix the pinch gesture in final project 1 by introducing an operation manager with a semaphore to serialize box attachments on ARKit, preventing multiple boxes and improving performance.

Learn to obtain a private api key for the inference api, create a token with read access, and safely use it in your app without exposing it publicly.

learn how to obtain a stability ai fast 3d api key and a second hanging face key, review pricing and credits, and copy the key into your app.

Create a new app called iArchitect that turns user input text into a 3d model available in iArchitect, with no storage or tests and a createKit repository.

Start by setting up the AR view in SwiftUI, create a session handler and ARViewContainer, configure AR world tracking, and fix entitlements with a camera usage description.

Introduce input view as a sheet with a text field and send button, show loading state with a circular progress view, and implement loadEntity to fetch an entity from input.

Introduce an entity service to load entities asynchronously and define an entity model with id and url. Inject the repository into the input view to enable AR container integration.

Explore loading and displaying a glb file in ARKit using the GLTFKit2 library, which translates GLB content into SceneKit nodes, and test with local data.

Introduce a file service by defining a save file method to write glb data to the documents directory, handle potential errors, and enable the entity repository to use this service.

Learn to implement a URLSessionHandler for performing requests, handle http responses, and fetch image data from a hugging face inference api using a dedicated image request builder and api key.

Implement a mesh repository that fetches a GLB mesh from Stability AI Fast3D using image data and saves it to local documents for ARKit.

Fine tunes the app by appending a white background suffix to user input before sending to stable diffusion. It tests red ball and green mushroom outputs to improve 3d rendering.