iOS Coming soon · preview

The iOS SDK is planned to share the Android SDK’s beauty pipeline, presets, parameter ranges and licences. Licences already accept iOS bundle IDs, so a licence issued today can list both your Android package and your iOS bundle ID.

Planned installation

Both Swift Package Manager and CocoaPods are planned.

// Package.swift (planned)
dependencies: [
    .package(url: "<Facevity iOS package URL>", from: "0.1.0")
]

# Podfile (planned)
pod 'Facevity'

Planned setup with a licence

import Facevity   // planned module name

let fv = try Facevity.initialize(options: .init(licenseServerURL: URL(string: "https://<licence server>")!))
fv.activateLicense(key: "FV-XXXX-XXXX-XXXX-XXXX") { info in
    print("Facevity licence:", info.status)
}
fv.setParams(.natural)          // .natural, .glam, .default, .none
fv.setSmoothIntensity(55)

As on Android, an offline licence file bundled with the app and a token from your own backend are planned alternatives.

Planned frame processing (CVPixelBuffer)

The iOS SDK is designed around CVPixelBuffer, which is what AVFoundation and most RTC SDKs deliver.

// Planned: processes the pixel buffer in place (BGRA or NV12) and returns true if it changed it.
let changed = fv.process(pixelBuffer, orientation: .up, mirrored: false, timestamp: time)

Planned AVFoundation integration

extension CameraController: AVCaptureVideoDataOutputSampleBufferDelegate {
    func captureOutput(_ output: AVCaptureOutput,
                       didOutput sampleBuffer: CMSampleBuffer,
                       from connection: AVCaptureConnection) {
        guard let pixelBuffer = CMSampleBufferGetImageBuffer(sampleBuffer) else { return }
        let time = CMSampleBufferGetPresentationTimeStamp(sampleBuffer)
        fv.process(pixelBuffer, orientation: .up, mirrored: false, timestamp: time)   // planned API
        // Hand the (processed) sample buffer to your preview, encoder or RTC SDK as before.
    }
}

Planned RTC hooks

The same rule as on Android applies: process the outgoing frame before the encoder, so remote participants receive the beautified video.

• Agora: a raw video frame observer on the capture position, processing the frame’s CVPixelBuffer in place.
• WebRTC / LiveKit: a capturer delegate that processes the RTCCVPixelBuffer before forwarding the frame to the video source.
• ZEGOCLOUD: custom video processing with a CVPixelBuffer buffer type.

Get early access

If iOS matters for your launch, tell us your timeline and the RTC SDK you use. Early-access partners help us prioritise the integrations above.