Federated Auth Realm Dart SDK

Official Dart SDK for interacting with the Savant Realms Federated Auth Realm API. It provides:

• A strongly typed FederatedAuthClient with helpers for registration, login, password flows, verification, and profile management
• Token storage abstractions with an in-memory implementation that works for server-side code
• Automatic bearer injection + refresh token retry logic with a session-expired callback
• A lightweight SRWP (Savant Realms Whispering Protocol) discovery client

Getting Started

Add the dependency to your Dart or Flutter project:

dart pub add federated_auth_realm_sdk_dart

Initialize the client

import 'package:federated_auth_realm_sdk_dart/federated_auth_realm_sdk_dart.dart';

final authClient = FederatedAuthClient(
  baseUrl: 'https://api.federatedauthrealm.com/v1',
  appKey: 'federated-auth-realm',
  onSessionExpired: () {
    // e.g. redirect user to login page
  },
);

Register or Login

final auth = await authClient.login(
  LoginRequest(email: 'user@example.com', password: 'secret'),
);

print('access token: \\${auth.accessToken}');
final profile = await authClient.getProfile();

Token helpers

final accessToken = await authClient.getAccessToken();
await authClient.setSessionTokens(
  accessToken: 'existing-access',
  refreshToken: 'existing-refresh',
);

SRWP discovery

final srwp = SRWPDiscoveryClient('https://api.federatedauthrealm.com/v1');
final knownRealms = await srwp.getKnownRealms();
await srwp.bootstrap('https://another-realm.example.com');

The SRWP client automatically normalizes the provided base URL so that /srwp endpoints are targeted even if you pass the public /v1 API URL.

Development

dart pub get
dart format .
dart test

Deployment & CI

• Use scripts/deploy.sh to run formatting, static analysis, unit tests, and dart pub publish --dry-run before optionally publishing (configure credentials via .env or environment variables).
• GitHub Actions (.github/workflows/ci.yml) keeps formatting/analyze/test checks green on pushes and pull requests, with a manual workflow option for a dry-run deploy.