LLM Insights on Swift

Platform-specific usage for the Swift SDK. For what LLM Insights are, the request/response model, required metrics, and the privacy notice, see the LLM Insights overview.

Insights summarize the buffered vitals, so breathing (the default set) and cardio must both be active. Assign both bundles to requestedMetrics:

sdk.config.requestedMetrics =
    SmartSpectraConfig.breathingMetrics + SmartSpectraConfig.cardioMetrics

cardioMetrics includes .arterialPressureTrace, which drives the on-screen pulse waveform. When requestedMetrics is left unset the SDK measures breathing only.

SmartSpectraSDK is @Observable (@MainActor). The latest insight is exposed as an observable property, updated on the main actor:

public internal(set) var insight: Insight?

Observe it like any other observable state — read it in a SwiftUI body, or react to changes:

.onChange(of: sdk.insight) { _, insight in
    guard let insight else { return }
    switch insight.result {
    case .analysis(let text): // the LLM text
        show(text)
    case .error:              // insight.error holds the message
        show("The insights service is currently unavailable.")
    case .none:
        break
    }
}

The same property delivers both the auto-fired periodic vitals insights and on-demand responses. Match insight.requestID against the value returned by requestInsight to tell them apart (auto-fired vitals won't match a request you made). Every insight is currently delivered with type == .vitals, so correlate on requestID, not type.

Call requestInsight on a running session (it throws otherwise). It returns the request ID used to correlate the asynchronous response:

@discardableResult
public func requestInsight(_ text: String) throws -> Int32

do {
    pendingRequestId = try sdk.requestInsight("Summarize my current vital signs and flag anything unusual.")
} catch {
    // e.g. processing not active
}

The prompt is combined with the latest buffered metrics when they exist, otherwise sent prompt-only.

Insight is a generated SwiftProtobuf struct. There are no hasAnalysis / hasError accessors — switch on the result oneof:

• insight.result → .analysis(String) on success, .error(String) on failure, or nil. (The convenience insight.analysis / insight.error properties return "" for the case that isn't set.)
• insight.requestID (Int32, note the capital ID) correlates the reply.
• insight.type is currently always .vitals (the .speech and .combined cases are reserved and not emitted today).

Full field documentation is in Data Types → Insight.

The first auto-fired insight arrives about 15 seconds after the session starts; allow that much valid measurement before an on-demand request can be grounded in the user's physiology.

• LLM Insights overview
• Swift API reference
• Data Types