Voice SDK for iOS

sent-dm-voice-swift lets an iOS app place Sent Voice calls: to a phone number, to another user of your app, or into a conference. Your backend mints a short-lived voice token for each user with the Sent API. The SDK registers the device with it, places calls, and reports how they go. What rings when a user calls is decided by your callback, as for every call; see Answering Calls.

Installation

A Swift package for iOS 15 or later. In Xcode, choose File → Add Package Dependencies and enter the repository URL, or add it to Package.swift:

dependencies: [
  .package(url: "https://github.com/sentdm/sent-dm-voice-swift.git", from: "0.1.0"),
],

Add the SentVoice product to your app target. Calls use the microphone, so add NSMicrophoneUsageDescription to the app's Info.plist; iOS asks the user on the first call.

Quick Start

// On your server, behind a route such as GET /api/voice-token
const response = await fetch('https://api.sent.dm/v3/voice/tokens', {
  method: 'POST',
  headers: { 'x-api-key': process.env.SENT_API_KEY, 'Content-Type': 'application/json' },
  body: JSON.stringify({ identity: user.id }),
});
const { data } = await response.json();
return data.token;

import SentVoice

let client = SentVoice.client(options: SentVoiceOptions(tokenProvider: {
  let (data, _) = try await URLSession.shared.data(from: URL(string: "https://example.com/api/voice-token")!)
  return String(decoding: data, as: UTF8.self)
}))

try await client.register()

let call = try await client.connect(to: "+14155551234")
let listener = call.on { event in
  switch event {
  case .connected: print("Connected")
  case .disconnected(let state, _): print("The call ended as \(state)")
  default: break
  }
}

Client Options

SentVoiceOptions takes the following options. Only tokenProvider is required.

Client

The client holds one registration and the call that runs on it.

Calls

connect(to:) and joinConference(name:) return a Call once the provider accepted it; the call then moves through ringing, answered and connected to completed, failed, busy or no_answer. One call at a time: a second connect while one is in progress throws CALL_IN_PROGRESS.

Errors

Every failure is a SentVoiceError with a code, a category, whether it is retriable, and a message you can show. The codes are the same on every platform:

Telemetry

The SDK reports call quality and errors to Sent so support can see what a user saw, batched and sent in the background. Set telemetryDisabled to turn it off. Nothing in the reports identifies the person beyond the identity your backend minted the token for.