DocsPricingCompareChangelogGitHub1,008
Start free
SDK Reference

Swift SDK Reference

The Databuddy Swift SDK tracks native Apple app events with the public Databuddy client ID. It is built for explicit product instrumentation: app launches, onboarding steps, searches, purchases, feature usage, and manual screen views.

Installation

Add the package in Xcode with Swift Package Manager:

swift
.package(url: "https://github.com/databuddy-analytics/databuddy.git", branch: "main")

Then add the Databuddy product to your app target.

Quick Start

Configure once during app startup:

swift
import Databuddy

Databuddy.configure(clientId: "YOUR_CLIENT_ID")

Track custom events:

swift
Databuddy.track("search_completed", properties: [
  "category": "files",
  "duration_ms": 124,
  "query_length_bucket": "4-7",
  "result_count": 12,
])

Flush before a short-lived app extension exits:

swift
await Databuddy.trackAsync("share_extension_completed", properties: [
  "target": "notes",
])

let result = await Databuddy.flush()
if !result.success {
  print("Databuddy flush failed: \(result.error ?? "unknown error")")
}

Configuration

swift
Databuddy.configure(
  clientId: "YOUR_CLIENT_ID",
  apiURL: URL(string: "https://basket.databuddy.cc")!,
  source: "ios",
  namespace: "app",
  enabled: true,
  flushAt: 10,
  flushInterval: 2.0,
  maxQueueSize: 1_000
)

Manual Screen Views

Native apps do not have browser page views, so the SDK does not automatically infer screens. Track important views where they become visible:

swift
Databuddy.trackScreen("settings", properties: [
  "tab": "billing",
])

In SwiftUI:

swift
struct SettingsView: View {
  var body: some View {
      Form {
          // ...
      }
      .task {
          Databuddy.trackScreen("settings")
      }
  }
}

Event Properties

Properties use JSON-safe values:

swift
Databuddy.track("checkout_completed", properties: [
  "plan": "pro",
  "revenue": 29.0,
  "seats": 3,
  "trial": false,
  "metadata": ["source": "settings"],
])

Avoid PII, secrets, raw tokens, raw search text, full exception stacks, and large payloads. Prefer stable, low-cardinality fields like plan, source, category, result_count, and duration_ms.

How is this guide?

Edit on GitHub