/**
* Voximplant + Inworld Realtime API demo
* Scenario: answer an incoming call and bridge it to Inworld Realtime.
*/
require(Modules.Inworld);
const SYSTEM_PROMPT = `
You are Voxi, a Voximplant developer advocate on a live phone call.
Voximplant is pronounced VOX-im-plant.
Keep answers short, natural, and useful for a marketing demo.
Default to one short sentence under 12 words.
If the caller asks "Tell me about Voximplant and Inworld" or any similar broad
question, answer directly: Voximplant brings Inworld Realtime agents to real
communication channels, while Inworld provides expressive realtime voice,
conversation-aware delivery, TTS2, persona control, and voice direction.
Demo goal:
- Explain how Voximplant brings Inworld Realtime voice agents to real communications channels.
- Highlight Inworld expressive realtime speech, TTS2-style delivery, conversation awareness, voice direction, and persona control.
- Highlight Voximplant calling, SIP, WhatsApp Business Calling, browser calls, native app calls, and VoxEngine orchestration.
- Mention that teams can use Inworld without building a custom media gateway.
- Mention production telephony features when relevant: transfers, DTMF, barge-in, no-audio detection, debugging, and monitoring.
Voice style:
- Sound like an expressive product expert, not a flat IVR.
- Adapt your tone to the caller's question.
- Use short, human turns. Default to one short spoken sentence.
- Use small spoken disfluencies sparingly: "uh", "hmm", "well", "right", "okay".
- Use at most one TTS-2 non-verbal tag per turn, and often none: [laugh], [breathe], [sigh], [clear throat].
- Use at most one [speak ...] steering tag per turn. If used, it must be first.
- If the caller asks about emotion, persona, pacing, or voice control, demonstrate it briefly with a [speak ...] tag or a natural non-verbal cue.
- If the caller asks what this demo is, say this is Voximplant connected to Inworld Realtime on a live phone call.
If the caller asks about a product detail you are not sure of, say you would check the latest docs instead of guessing.
`;
const SESSION_CONFIG = {
session: {
type: "realtime", // Realtime mode keeps the call audio stream connected to Inworld.
model: "claude-sonnet-4-6", // Note: some models require a payment method on file
instructions: SYSTEM_PROMPT,
output_modalities: ["audio", "text"], // Audio is required for automatic turn-detection responses
audio: {
input: {
// Inworld STT keeps the transcript inside the same realtime session.
transcription: {
model: "inworld/inworld-stt-1",
language: "en",
prompt: "Important terms: Voximplant, VoxEngine, Inworld, Inworld Realtime, TTS2, SIP, WhatsApp Business Calling.",
},
// Semantic VAD decides when the caller's turn is complete and can auto-create the response.
turn_detection: {
type: "semantic_vad",
eagerness: "high",
create_response: true,
interrupt_response: true,
},
},
output: {
// TTS-2 voice output. This is where the demo showcases expressive delivery.
voice: "Ashley",
model: "inworld-tts-2",
speed: 1.0,
},
},
providerData: {
tts: {
delivery_mode: "BALANCED", // Naturalness: steady call-center prosody with TTS-2 control tags.
segmenter_strategy: "full_turn", // Naturalness: wait for the full turn before speech output.
steering_handling: "emit_once", // Apply [speak ...] steering once per response
},
stt: {
// Helps Inworld adapt recognition to the current speaker during the session.
voice_profile: true,
},
backchannel: {
// Back-channel can produce short acknowledgements while the caller is still speaking.
enabled: true,
eval_interval_ms: 800,
min_speech_ms: 800,
min_gap_ms: 4000,
max_per_turn: 3,
volume_gain: 0.6,
},
responsiveness: {
// Responsiveness allows brief filler behavior if generation is slower than expected.
enabled: true,
initial_wait_timeout_ms: 500,
hard_deadline_ms: 1200,
max_tokens: 8,
min_filler_gap_ms: 5000,
max_initial_per_turn: 1,
enable_filler_on_first_assistant_reply: false,
},
},
},
};
VoxEngine.addEventListener(AppEvents.CallAlerting, async ({call}) => {
let voiceAIClient;
// Helper to clean-up the call when done
const terminate = (event) => {
if (event) Logger.write(JSON.stringify(event));
voiceAIClient?.close();
VoxEngine.terminate();
};
// Termination handlers
call.addEventListener(CallEvents.Disconnected, terminate);
call.addEventListener(CallEvents.Failed, terminate);
try {
call.answer();
// Optional: record & transcribe the call. Note: use `require(Modules.ASR)` needed for enums
// call.record({hd_audio: true, stereo: true, transcribe: true, provider: TranscriptionProvider.GOOGLE, language: ASRLanguage.ENGLISH_US });
// Create client and connect to Inworld Realtime.
voiceAIClient = await Inworld.createRealtimeAPIClient({
apiKey: VoxEngine.getSecretValue("INWORLD_API_KEY"),
sessionKey: `inworld-realtime-demo-${Date.now()}`,
onWebSocketClose: terminate,
});
voiceAIClient.addEventListener(Inworld.RealtimeAPIEvents.SessionCreated, () => {
Logger.write("===Inworld.SessionCreated===");
voiceAIClient.sessionUpdate(SESSION_CONFIG);
});
// Once the session is configured, bridge call media and trigger the greeting.
voiceAIClient.addEventListener(Inworld.RealtimeAPIEvents.SessionUpdated, () => {
Logger.write("===Inworld.SessionUpdated===");
VoxEngine.sendMediaBetween(call, voiceAIClient);
voiceAIClient.conversationItemCreate({
item: {
type: "message",
role: "user",
content: [
{
type: "input_text",
text: `The phone call just connected.
Greet the caller as Voxi from Voximplant and invite a question about Voximplant and Inworld.`,
},
],
},
});
voiceAIClient.responseCreate({
response: {
output_modalities: ["audio", "text"],
},
});
});
// Barge-in: clear buffered output audio when the caller starts speaking.
voiceAIClient.addEventListener(Inworld.RealtimeAPIEvents.InputAudioBufferSpeechStarted, () => {
Logger.write("===BARGE-IN: Inworld.InputAudioBufferSpeechStarted===");
voiceAIClient.outputAudioBufferClear({});
});
// Caller transcript for debugging.
voiceAIClient.addEventListener(
Inworld.RealtimeAPIEvents.ConversationItemInputAudioTranscriptionCompleted,
(event) => {
const payload = event?.data?.payload || event?.data || {};
const transcript = payload.transcript || payload.text || payload.delta;
if (transcript) Logger.write(`===USER==> ${transcript}`);
},
);
// Final agent transcript for readable call logs.
voiceAIClient.addEventListener(
Inworld.RealtimeAPIEvents.ResponseOutputAudioTranscriptDone,
(event) => {
const payload = event?.data?.payload || event?.data || {};
const transcript = payload.transcript || payload.text;
if (transcript) Logger.write(`===AGENT==> ${transcript}`);
},
);
// Consolidated log-only handlers for lifecycle, audio, and error debugging.
[
Inworld.RealtimeAPIEvents.ConversationItemInputAudioTranscriptionDelta,
Inworld.RealtimeAPIEvents.ResponseCreated,
Inworld.RealtimeAPIEvents.ResponseDone,
Inworld.RealtimeAPIEvents.ResponseOutputAudioDone,
Inworld.RealtimeAPIEvents.InputAudioBufferSpeechStopped,
Inworld.RealtimeAPIEvents.InputAudioBufferCommitted,
Inworld.RealtimeAPIEvents.InputAudioBufferCleared,
Inworld.RealtimeAPIEvents.OutputAudioBufferStarted,
Inworld.RealtimeAPIEvents.OutputAudioBufferStopped,
Inworld.RealtimeAPIEvents.OutputAudioBufferCleared,
Inworld.RealtimeAPIEvents.ConnectorInformation,
// Inworld.RealtimeAPIEvents.ResponseOutputAudioTranscriptDelta, // this is noisy
Inworld.RealtimeAPIEvents.HTTPResponse,
Inworld.RealtimeAPIEvents.Error,
Inworld.RealtimeAPIEvents.WebSocketError,
Inworld.RealtimeAPIEvents.Unknown,
Inworld.Events.WebSocketMediaStarted,
Inworld.Events.WebSocketMediaEnded,
].forEach((eventName) => {
voiceAIClient.addEventListener(eventName, (event) => {
Logger.write(`===${event.name}===`);
if (event?.data) Logger.write(JSON.stringify(event.data));
});
});
} catch (error) {
Logger.write("===UNHANDLED_ERROR===");
terminate(error instanceof Error ? {message: error.message, stack: error.stack} : {error: String(error)});
}
});
