Photokit logo

Photokit

CommunityPopular
dpearson2699
photokit

Implement, review, or improve photo picking, camera capture, and media handling in iOS apps using PhotoKit and AVFoundation. Use when working with PhotosPicker, PHPickerViewController, camera capture sessions (AVCaptureSession), photo library access, image loading and display, video recording, or media permissions. Also use when selecting photos from the library, taking pictures, recording video, processing images, or handling photo/camera privacy permissions in Swift apps.

Overview

Publisherdpearson2699
Repositoryswift-ios-skills
Skill namephotokit
Stars
1.2K
Forks
61
Bundled files
5
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 5 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by dpearson2699 on GitHub. Read the source before you install it.

Installation

Install the Photokit AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/dpearson2699/swift-ios-skills/tree/main/skills/photokit
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/dpearson2699/swift-ios-skills.git /tmp/swift-ios-skills
mkdir -p .claude/skills
cp -r /tmp/swift-ios-skills/skills/photokit .claude/skills/photokit
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Photokit in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Photokit on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Photokit is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

PhotoKit

Modern patterns for photo picking, camera capture, image loading, and media permissions targeting iOS 26+ with Swift 6.3. Patterns are backward-compatible to iOS 16 unless noted. See references/photokit-patterns.md for complete picker recipes and references/camera-capture.md for AVCaptureSession patterns.

Contents

PhotosPicker (SwiftUI, iOS 16+)

PhotosPicker is the native SwiftUI replacement for UIImagePickerController. It runs out-of-process, requires no photo library permission for browsing, and supports single or multi-selection with media type filtering.

Single Selection

swift
import SwiftUI
import PhotosUI

struct SinglePhotoPicker: View {
    @State private var selectedItem: PhotosPickerItem?
    @State private var selectedImage: Image?

    var body: some View {
        VStack {
            if let selectedImage {
                selectedImage
                    .resizable()
                    .scaledToFit()
                    .frame(maxHeight: 300)
            }

            PhotosPicker("Select Photo", selection: $selectedItem, matching: .images)
        }
        .onChange(of: selectedItem) { _, newItem in
            Task {
                if let data = try? await newItem?.loadTransferable(type: Data.self),
                   let uiImage = UIImage(data: data) {
                    selectedImage = Image(uiImage: uiImage)
                }
            }
        }
    }
}

Multi-Selection

swift
struct MultiPhotoPicker: View {
    @State private var selectedItems: [PhotosPickerItem] = []
    @State private var selectedImages: [Image] = []

    var body: some View {
        VStack {
            ScrollView(.horizontal) {
                HStack {
                    ForEach(selectedImages.indices, id: \.self) { index in
                        selectedImages[index]
                            .resizable()
                            .scaledToFill()
                            .frame(width: 100, height: 100)
                            .clipShape(.rect(cornerRadius: 8))
                    }
                }
            }

            PhotosPicker(
                "Select Photos",
                selection: $selectedItems,
                maxSelectionCount: 5,
                matching: .images
            )
        }
        .onChange(of: selectedItems) { _, newItems in
            Task {
                selectedImages = []
                for item in newItems {
                    if let data = try? await item.loadTransferable(type: Data.self),
                       let uiImage = UIImage(data: data) {
                        selectedImages.append(Image(uiImage: uiImage))
                    }
                }
            }
        }
    }
}

Media Type Filtering

Filter with PHPickerFilter composites to restrict selectable media:

swift
// Images only
PhotosPicker(selection: $items, matching: .images)

// Videos only
PhotosPicker(selection: $items, matching: .videos)

// Live Photos only
PhotosPicker(selection: $items, matching: .livePhotos)

// Screenshots only
PhotosPicker(selection: $items, matching: .screenshots)

// Images and videos combined
PhotosPicker(selection: $items, matching: .any(of: [.images, .videos]))

// Images excluding screenshots
PhotosPicker(selection: $items, matching: .all(of: [.images, .not(.screenshots)]))

Loading Selected Items with Transferable

PhotosPickerItem loads content asynchronously via loadTransferable(type:). Define a Transferable type for automatic decoding:

swift
struct PickedImage: Transferable {
    let data: Data
    let image: Image

    static var transferRepresentation: some TransferRepresentation {
        DataRepresentation(importedContentType: .image) { data in
            guard let uiImage = UIImage(data: data) else {
                throw TransferError.importFailed
            }
            return PickedImage(data: data, image: Image(uiImage: uiImage))
        }
    }
}

enum TransferError: Error {
    case importFailed
}

// Usage
if let picked = try? await item.loadTransferable(type: PickedImage.self) {
    selectedImage = picked.image
}

Always load in a Task to avoid blocking the main thread. Handle nil returns and thrown errors -- the user may select a format that cannot be decoded.

Privacy and Permissions

Photo Library Access Levels

iOS provides two access levels for the photo library. The system automatically presents the limited-library picker when an app requests .readWrite access -- users choose which photos to share.

Access LevelDescriptionInfo.plist Key
Add-onlyWrite photos to the library without readingNSPhotoLibraryAddUsageDescription
Read-writeFull or limited read access plus writeNSPhotoLibraryUsageDescription

PhotosPicker requires no permission to browse -- it runs out-of-process and only grants access to selected items. Request explicit permission only when you need to read the full library (e.g., a custom gallery) or save photos.

Checking and Requesting Photo Library Permission

swift
import Photos

func requestPhotoLibraryAccess() async -> PHAuthorizationStatus {
    let status = PHPhotoLibrary.authorizationStatus(for: .readWrite)

    switch status {
    case .notDetermined:
        return await PHPhotoLibrary.requestAuthorization(for: .readWrite)
    case .authorized, .limited:
        return status
    case .denied, .restricted:
        return status
    @unknown default:
        return status
    }
}

Camera Permission

Add NSCameraUsageDescription to Info.plist. Check and request access before configuring a capture session:

swift
import AVFoundation

func requestCameraAccess() async -> Bool {
    let status = AVCaptureDevice.authorizationStatus(for: .video)

    switch status {
    case .notDetermined:
        return await AVCaptureDevice.requestAccess(for: .video)
    case .authorized:
        return true
    case .denied, .restricted:
        return false
    @unknown default:
        return false
    }
}

Handling Denied Permissions

When the user denies access, guide them to Settings. Never repeatedly prompt or hide functionality silently.

swift
struct PermissionDeniedView: View {
    let message: String
    @Environment(\.openURL) private var openURL

    var body: some View {
        ContentUnavailableView {
            Label("Access Denied", systemImage: "lock.shield")
        } description: {
            Text(message)
        } actions: {
            Button("Open Settings") {
                if let url = URL(string: UIApplication.openSettingsURLString) {
                    openURL(url)
                }
            }
        }
    }
}

Required Info.plist Keys

KeyWhen Required
NSPhotoLibraryUsageDescriptionReading photos from the library
NSPhotoLibraryAddUsageDescriptionSaving photos/videos to the library
NSCameraUsageDescriptionAccessing the camera
NSMicrophoneUsageDescriptionRecording audio (video with sound)

Omitting a required key causes a runtime crash when the permission dialog would appear.

Camera Capture Basics

Own each capture session in a dedicated controller and serialize configuration, startRunning(), and stopRunning() on the same non-main executor. Never mix main-actor configuration with detached start/stop tasks: beginConfiguration()/commitConfiguration() and session state changes must not race. The representable view only displays the preview.

Minimal Camera Manager

Load Camera Capture for the serialized-controller pattern, photo/video delegates, focus, torch, orientation, and scanning. The critical lifecycle is:

  1. Request access outside the session configuration transaction.
  2. On the capture executor, call beginConfiguration() and immediately install defer { commitConfiguration() } so every early exit balances the transaction.
  3. Add inputs and outputs only after canAddInput/canAddOutput checks.
  4. Start or stop on that same executor, then publish UI state on the main actor only after the synchronous call returns.
  5. On failure, stop, restore a fresh session fixture, fix the configuration, and rerun authorization, background/foreground, interruption, and capture checks.

Camera Preview in SwiftUI

Wrap AVCaptureVideoPreviewLayer in a UIViewRepresentable. Override layerClass for automatic resizing:

swift
import SwiftUI
import AVFoundation

struct CameraPreview: UIViewRepresentable {
    let session: AVCaptureSession

    func makeUIView(context: Context) -> PreviewView {
        let view = PreviewView()
        view.previewLayer.session = session
        view.previewLayer.videoGravity = .resizeAspectFill
        return view
    }

    func updateUIView(_ uiView: PreviewView, context: Context) {
        if uiView.previewLayer.session !== session {
            uiView.previewLayer.session = session
        }
    }
}

final class PreviewView: UIView {
    override class var layerClass: AnyClass { AVCaptureVideoPreviewLayer.self }
    var previewLayer: AVCaptureVideoPreviewLayer { layer as! AVCaptureVideoPreviewLayer }
}

Using the Camera in a View

swift
struct CameraScreen: View {
    @State private var cameraManager = CameraManager()

    var body: some View {
        ZStack(alignment: .bottom) {
            CameraPreview(session: cameraManager.session)
                .ignoresSafeArea()

            Button {
                // Capture photo -- see references/camera-capture.md
            } label: {
                Circle()
                    .fill(.white)
                    .frame(width: 72, height: 72)
                    .overlay(Circle().stroke(.gray, lineWidth: 3))
            }
            .padding(.bottom)
        }
        .task {
            await cameraManager.configure()
            cameraManager.start()
        }
        .onDisappear {
            cameraManager.stop()
        }
    }
}

Always call stop() in onDisappear. A running capture session holds the camera exclusively and drains battery.

Image Loading and Display

AsyncImage for Remote Images

swift
AsyncImage(url: imageURL) { phase in
    switch phase {
    case .empty:
        ProgressView()
    case .success(let image):
        image
            .resizable()
            .scaledToFill()
    case .failure:
        Image(systemName: "photo")
            .foregroundStyle(.secondary)
    @unknown default:
        EmptyView()
    }
}
.frame(width: 200, height: 200)
.clipShape(.rect(cornerRadius: 12))

AsyncImage does not cache images across view redraws. For production apps with many images, use a dedicated image loading library or URLCache-based caching.

Downsampling Large Images

Load full-resolution photos from the library into a display-sized CGImage to avoid memory spikes. A 48MP photo can consume over 200 MB uncompressed.

swift
import ImageIO
import UIKit

func downsample(data: Data, to pointSize: CGSize, scale: CGFloat = UITraitCollection.current.displayScale) -> UIImage? {
    let maxDimensionInPixels = max(pointSize.width, pointSize.height) * scale

    let options: [CFString: Any] = [
        kCGImageSourceCreateThumbnailFromImageAlways: true,
        kCGImageSourceShouldCacheImmediately: true,
        kCGImageSourceCreateThumbnailWithTransform: true,
        kCGImageSourceThumbnailMaxPixelSize: maxDimensionInPixels
    ]

    guard let source = CGImageSourceCreateWithData(data as CFData, nil),
          let cgImage = CGImageSourceCreateThumbnailAtIndex(source, 0, options as CFDictionary) else {
        return nil
    }

    return UIImage(cgImage: cgImage)
}

Use this whenever displaying user-selected photos in lists, grids, or thumbnails. Pass the raw Data from PhotosPickerItem directly to the downsampler before creating a UIImage.

Image Rendering Modes

swift
// Original: display the image as-is with its original colors
Image("photo")
    .renderingMode(.original)

// Template: treat the image as a mask, colored by foregroundStyle
Image(systemName: "heart.fill")
    .renderingMode(.template)
    .foregroundStyle(.red)

Use .original for photos and artwork. Use .template for icons that should adopt the current tint color.

Common Mistakes

DON'T: Use UIImagePickerController for photo picking. DO: Use PhotosPicker (SwiftUI) or PHPickerViewController (UIKit). Why: UIImagePickerController is legacy API with limited functionality. PhotosPicker runs out-of-process, supports multi-selection, and requires no library permission for browsing.

DON'T: Request full photo library access when you only need the user to pick photos. DO: Use PhotosPicker which requires no permission, or request .readWrite and let the system handle limited access. Why: Full access is unnecessary for most pick-and-use workflows. The system's limited-library picker respects user privacy and still grants access to selected items.

DON'T: Load full-resolution images into memory for thumbnails. DO: Use CGImageSource with kCGImageSourceThumbnailMaxPixelSize to downsample. A 48MP image is over 200 MB uncompressed.

DON'T: Block the main thread loading PhotosPickerItem data. DO: Use async loadTransferable(type:) in a Task.

DON'T: Forget to stop AVCaptureSession when the view disappears. DO: Call session.stopRunning() in onDisappear or dismantleUIView.

DON'T: Assume camera access is granted without checking. DO: Check AVCaptureDevice.authorizationStatus(for: .video) and handle .denied/.restricted.

DON'T: Call session.startRunning() on the main thread. DO: Run it on the same dedicated serial executor that owns configuration and stop operations. Why: startRunning() is a synchronous blocking call that can take hundreds of milliseconds while the hardware initializes.

DON'T: Create AVCaptureSession inside a UIViewRepresentable. DO: Own the session in a separate @Observable model.

Review Checklist

  • PhotosPicker used instead of deprecated UIImagePickerController
  • Privacy descriptions in Info.plist for camera/photo library
  • Loading states handled for async image/video loading
  • Large images downsampled with CGImageSource before display
  • Camera session started on background thread; stopped in onDisappear
  • Permission denial handled with Settings deep link
  • AVCaptureSession owned by model, not created inside UIViewRepresentable
  • Media asset types and picker results are Sendable across concurrency boundaries

References

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Photokit AI skill do?

Implement, review, or improve photo picking, camera capture, and media handling in iOS apps using PhotoKit and AVFoundation. Use when working with PhotosPicker, PHPickerViewController, camera capture sessions (AVCaptureSession), photo library access, image loading and display, video recording, or media permissions. Also use when selecting photos from the library, taking pictures, recording video, processing images, or handling photo/camera privacy permissions in Swift apps.

Why use Photokit on TypingMind?

Because you install it once and use it with any model. Photokit is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Photokit in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/dpearson2699/swift-ios-skills/tree/main/skills/photokit. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Photokit?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Photokit?

As many as you like. As long as a model supports skills, you can use Photokit with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Photokit AI skill free?

It is published on GitHub by dpearson2699. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇