kletso_voice #

Microphone and speaker for Kletso voice sessions. package:kletso_flutter draws the mic button, the voice sheet and the avatar; this package adds the audio and installs itself into KletsoClient in one line.

import 'package:kletso_core/kletso_core.dart';
import 'package:kletso_voice/kletso_voice.dart';

final client = await Kletso.init(config);
KletsoVoice.install(client);   // capture + playback for client.voice

After that, KletsoChat shows the mic button whenever the agent has voice enabled in the Kletso dashboard, and client.voice (a KletsoVoiceController) can be driven directly: start(), stop(), commit() for push-to-talk, state, captions and interruption events.

What it does #

• Audio travels as binary PCM frames on the same kletso.v1 WebSocket as the chat (KletsoAudioFrame from kletso_ui_schema).
• A level meter feeds the avatar lip-sync in kletso_flutter.
• KletsoDeviceAudioIo is the default KletsoAudioIo; hosts with their own audio stack implement KletsoAudioIo from package:kletso_core instead and skip this package.

Permissions #

• Android: RECORD_AUDIO in AndroidManifest.xml.
• iOS / macOS: NSMicrophoneUsageDescription in Info.plist (macOS also needs the audio-input entitlement).
• Web: the browser asks on first use; serve over https.

KletsoVoice.requestPermission() asks ahead of time; the controller asks on start() otherwise.

Low-level pieces #

• KletsoCapture / KletsoCaptureConfig: the platform capture interface (requestPermission, start → Stream<Uint8List> of PCM16, stop, dispose).
• KletsoDeviceAudioIo, KletsoPcmSink: capture + playback bundle used by install.

See also #

• Docs: Voice and avatar
• kletso_flutter: chat UI, voice sheet, avatar
• kletso_core: client, KletsoVoiceController, KletsoAudioIo