arcane_voice 1.0.0

Flutter client runtime for provider-agnostic realtime voice calls through an Arcane Voice proxy.

arcane_voice

arcane_voice is the Flutter client runtime for Arcane Voice realtime calls. It owns the proxy-facing websocket transport, microphone capture, streamed audio playback, call/session orchestration, and transcript state management.

This package is intentionally UI-light. It gives apps the call runtime and state; the demo UI lives in the repository example app.

Public surface

• CallSessionController Coordinates the websocket session, audio services, provider selection, voice selection, mute state, and transcript state.
• RealtimeSocketClient Sends and receives the typed realtime protocol over websocket, with audio kept as binary frames.
• AudioCaptureService Captures microphone PCM audio suitable for proxy streaming.
• AudioPlaybackService Buffers and plays streamed PCM audio replies.
• TranscriptTimeline Reduces transcript delta/final events into stable conversation history.
• shared exports from arcane_voice_models Provider catalog, protocol models, and turn-detection config are re-exported so client apps do not need to import both packages directly.

Getting started

Point the controller at a running Arcane Voice proxy:

import 'package:arcane_voice/arcane_voice.dart';

CallSessionController controller = CallSessionController(
  serverUrl: 'ws://127.0.0.1:8080/ws/realtime',
);

Typical app flow:

await controller.startCall();

controller.onProviderChanged(RealtimeProviderCatalog.gemini);
controller.onVoiceChanged('Kore');

await controller.stopCall();
controller.dispose();

The controller exposes the current:

• session state
• transcript entries
• provider, model, and voice selection
• start, stop, mute, and unmute actions
• latest proxy/provider error

Permissions

Apps using this package need microphone and network permissions. The repository example app includes working Android, iOS, and macOS setup.

Protocol boundary

Structured control/config traffic uses the typed models from arcane_voice_models. Audio remains raw binary websocket frames so the client runtime can stay low-latency and provider-agnostic.

Related packages

• arcane_voice_models Shared protocol models and codec
• arcane_voice_proxy Proxy server runtime