voice-cat/clients/apple/dotnet/README.md

Managed Apple clients

VoiceCat.Mac and VoiceCat.iOS are the native AppKit and UIKit C# clients. They target the .NET 10 Apple workloads and reference the same VoiceCat.Core and VoiceCat.Audio assemblies used by Windows and the managed CLI. UIKit was selected over MAUI to retain direct AVAudioSession/AVAudioEngine lifecycle control and native VoiceOver behavior without another UI abstraction.

The managed client now implements the Swift client's functional surface: profiles and Keychain authentication, TOFU, protected channels, hierarchical channel presentation and roster state, channel and modeless private text, microphone/auxiliary/screen-audio streams, selectable Core Audio devices, VAD/PTT/always-on input, stereo microphone, RNNoise, per-stream receive tuning, self/server mute and deafen, full channel configuration, moderation, permissions, account administration, event sounds and speech. It imports the legacy Swift profile, TOFU and Keychain state during cutover. Capture is converted to interleaved 48 kHz int16 PCM through AVAudioConverter; playback converts the shared bounded PcmRing into Core Audio's native planar Float32 layout inside an allocation-free, non-blocking AVAudioSourceNode callback.

The Swift app remains the release oracle until the managed app passes the manual VoiceOver and real multi-human call matrix. Produce and validate an ad-hoc Release bundle with zsh clients/apple/dotnet/publish-macos.sh --dry-run. The script prefers /usr/local/share/dotnet/dotnet, where the pinned Apple workload is installed; set VOICECAT_DOTNET to override that host. Signing is performed on bin/Release/distribution/VoiceCat.app, leaving MSBuild's incremental app bundle untouched. For distribution, set VOICECAT_CODESIGN_IDENTITY; setting APPLE_ID, APPLE_TEAM_ID, and APPLE_APP_PASSWORD additionally submits, staples, and Gatekeeper-validates the notarized bundle.

Build on Apple Silicon macOS 27 with Xcode 27, .NET SDK 10.0.401 and workload set 10.0.401. Homebrew protobuf supplies a native arm64 protoc; the current Grpc.Tools package contains only an x64 macOS compiler.

sudo dotnet workload install macos ios --version 10.0.401
brew install protobuf # if /opt/homebrew/bin/protoc is not already present
cmake -S dotnet/native -B dotnet/artifacts/native-build -DCMAKE_BUILD_TYPE=Release -DVOICECAT_DOTNET_RID=osx-arm64
cmake --build dotnet/artifacts/native-build --config Release --target voicecat_media --parallel 2
cmake --install dotnet/artifacts/native-build --config Release --component DotnetMedia --prefix dotnet/artifacts/native
dotnet restore clients/apple/dotnet/VoiceCat.Apple.slnx
dotnet build clients/apple/dotnet/VoiceCat.Apple.slnx -c Debug
open clients/apple/dotnet/VoiceCat.Mac/bin/Debug/net10.0-macos27.0/osx-arm64/VoiceCat.app

The iOS build first stages static device and simulator Opus/RNNoise archives, then builds the UIKit host. MSBuild also builds and embeds the existing Swift ReplayKit upload extension; that deliberately remains Swift because of the extension's tight memory budget and unsupported managed extension runtime.

./dotnet/build-native-ios.sh
dotnet build clients/apple/dotnet/VoiceCat.iOS/VoiceCat.iOS.csproj -c Debug -r iossimulator-arm64

For a physical device, use the checked-in build and deployment wrappers. The iPhone must be paired and trusted, and Xcode must have an Apple Development identity and provisioning profile for both me.iamtalon.voicecat and me.iamtalon.voicecat.broadcast. Let automatic signing select them, or set VOICECAT_CODESIGN_KEY and VOICECAT_CODESIGN_PROVISION before building.

clients/apple/dotnet/build-ios-device.sh --configuration Debug
clients/apple/dotnet/deploy-ios-device.sh --list
clients/apple/dotnet/deploy-ios-device.sh --device "My iPhone" --configuration Debug --console

The build stages the verified app at dist/ios-managed-device/VoiceCat.iOS.app. Pass --no-build to the deployment script for quick reinstall cycles. On iOS 18–26, screen audio uses the retained ReplayKit extension. On iOS 27 and newer, the host uses a small dynamically loaded ScreenCaptureKit bridge and writes the same versioned ring; the Settings row toggles that capture on and off. This keeps one managed consumer and allows the app to remain launchable on older supported systems.

While joined to voice, the active PlayAndRecord session and audio engine remain running when the scene backgrounds or the device locks, which keeps microphone capture, peer playback and the screen-audio pump eligible for the declared audio background mode. Foreground activation, hardware route changes, audio interruptions and media-service resets revalidate or rebuild the audio graph. Validate this behavior on hardware: iOS simulator lifecycle transitions do not prove background execution, lock-screen routing or Bluetooth recovery.

Profiles, TOFU state, passwords and the ReplayKit ring use the signed App Group group.me.iamtalon.voicecat; the managed client migrates the old app-private profile files on first use. The shared ring ABI is frozen in docs/broadcast-ring-format.md.

The native build stages an osx-arm64 libvoicecat_media.dylib; that shim contains only Opus/RNNoise. Debug builds deliberately omit hardened runtime so an ad-hoc-signed local app can load the separately ad-hoc-signed .NET runtime libraries without an Apple Development identity. Release builds retain hardened runtime for Developer ID signing and notarization. Only a macOS host can link, launch, grant microphone access and verify live devices. Final audio quality is validated with real multi-human calls after the feature surface is complete; a synthetic ten-minute sine-wave listen is deliberately not a release gate.