Remote Config Reference

Value Accessors

Every SDK matches the Firebase SDK of its own platform: Firebase Android naming on Android and KMP, a native Swift API on iOS, and the familiar platform shapes on Flutter, React Native, Unity, and Web. Values resolve remote → in-app default → static zero-value, and accessors never block the calling thread.

Custom Signals

Custom signals let you target config values based on app-specific data. Unlike Firebase, signals never leave the device because evaluation is local.

val signals = CustomSignals.Builder()
    .put("player_level", 42L)
    .put("subscription_tier", "gold")
    .build()
// suspend function, call from a coroutine
AppSpikeRemoteConfig.setCustomSignals(signals)

try await AppSpike.remoteConfig().setCustomSignals([
    "player_level": 42,
    "subscription_tier": "gold",
])

val signals = CustomSignals.Builder()
    .put("player_level", 42L)
    .put("subscription_tier", "gold")
    .build()
// suspend function, call from a coroutine
AppSpikeRemoteConfig.setCustomSignals(signals)

await remoteConfig.setCustomSignals({
  'player_level': 42,
  'subscription_tier': 'gold',
});

await remoteConfig().setCustomSignals({ player_level: 42, subscription_tier: 'gold' });

await remoteConfig.SetCustomSignalsAsync(new CustomSignals.Builder()
    .Put("player_level", 42L)
    .Put("subscription_tier", "gold")
    .Build());

await setCustomSignals(remoteConfig, { player_level: 42, subscription_tier: 'gold' });

Value Types

Four types: string, boolean, number, json. Type is fixed once a parameter is published, so changing it requires delete + recreate. JSON has no dedicated accessor on Android: read with getString(key) and parse with your own tooling (same as Firebase). On iOS, JSON parameters decode directly to Codable types via the subscript API.

Targeting Conditions

Conditions are named, reusable predicates. Each condition is a list of clauses combined with AND. For each parameter, the first matching condition (top-down) wins, same as Firebase.

Percentage Rollouts

Each device is bucketed deterministically by SHA-256 of device ID + seed, with 4-decimal precision. Clauses that share a seed name divide the same set of users, so you can build groups that never overlap for controlled experiments. between 0 and 50 and between 50 and 100 on the same seed never overlap. A device keeps its bucket across fetches.

Limits