Call Intelligence API

Modes

• agent_assist — human agent stays on the call; Lugemi captions + suggests replies.
• ivr_bot — full auto voice bot (STT → chat → TTS).

1 · Create session

curl -sS -X POST "$LUGEMI_API_URL/v1/call-intelligence/sessions" \
  -H "Authorization: Bearer $LUGEMI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"mode":"agent_assist","consentAttested":true,"agentLanguage":"en"}'

2 · Turn

SESSION=cis_…
curl -sS -X POST "$LUGEMI_API_URL/v1/call-intelligence/sessions/$SESSION/turn" \
  -H "Authorization: Bearer $LUGEMI_API_KEY" \
  -F "file=@caller.wav" \
  -F "speakToCaller=false"

3 · TypeScript SDK

import { Lugemi } from '@lugemi/sdk';

const client = new Lugemi({ apiKey: process.env.LUGEMI_API_KEY! });
const { session } = await client.callIntelligence.createSession({
  mode: 'ivr_bot',
  consentAttested: true,
});
const turn = await client.callIntelligence.turn({
  sessionId: session.id as string,
  text: 'Hello, I need help in Twi.',
  speakToCaller: true,
});

Package: @lugemi/sdk · Mobile: @lugemi/sdk-react-native · iOS packages/sdk-ios · Android packages/sdk-android

4 · React Native

import { Lugemi, appendMobileAudio, recordingFileFromUri } from '@lugemi/sdk-react-native';
// Record with expo-av / react-native-audio-api, then:
const form = new FormData();
appendMobileAudio(form, recordingFileFromUri({ uri }));
await fetch(`${API}/v1/call-intelligence/sessions/${id}/turn`, {
  method: 'POST',
  headers: { Authorization: `Bearer ${KEY}` },
  body: form,
});

Bridges

• Twilio POST /v1/call-intelligence/bridges/twilio/voice
• LiveKit POST /v1/call-intelligence/bridges/livekit/webhook
• Amazon Connect POST /v1/call-intelligence/bridges/amazon-connect/lambda
• Genesys POST /v1/call-intelligence/bridges/genesys/architect
• Vonage / Telnyx HTTP event bridges
• WebRTC offer POST /v1/call-intelligence/ingress/webrtc/offer
• SIP trunks POST /v1/call-intelligence/ingress/sip/trunks

Catalog (no auth): GET /v1/call-intelligence/engine