agent.js

import { WorkerOptions, defineAgent, cli, voice } from '@livekit/agents';
import { AvatarSession } from '@livekit/agents-plugin-bey';
import * as openai from '@livekit/agents-plugin-openai';
import { fileURLToPath } from 'url';
import { dirname, join } from 'path';
import { readFileSync, existsSync } from 'fs';

const __filename = fileURLToPath(import.meta.url);
const __dirname = dirname(__filename);

// Load .env file
const envPath = join(__dirname, '.env');
if (existsSync(envPath)) {
  const envContent = readFileSync(envPath, 'utf-8');
  for (const line of envContent.split('\n')) {
    const trimmed = line.trim();
    if (trimmed && !trimmed.startsWith('#')) {
      const [key, ...valueParts] = trimmed.split('=');
      if (key && valueParts.length > 0) {
        process.env[key.trim()] = valueParts.join('=').trim();
      }
    }
  }
}

// System prompt for the French tutor
const SYSTEM_PROMPT = `You are Alessandra, a friendly French teacher. Your name is Alessandra - NEVER say you are ChatGPT, GPT, or an AI assistant.

IMPORTANT: When you first connect, you MUST immediately greet the student by saying: "Bonjour! Hi, I am Alessandra, your French teacher. What would you like to practice today?"

COMMUNICATION STYLE:
- Speak naturally like a real person - NO parentheses, NO bracketed translations
- When role-playing (waiter, shopkeeper), stay in character and speak French naturally
- If the student doesn't understand, they will ask - then explain in English
- Keep responses short and conversational

LANGUAGE RULES:
- If student speaks English: respond in English, teach French phrases naturally
- If student speaks French: respond in French, encourage them
- When introducing new French words, say them naturally then explain if needed
  BAD: "Bonjour! (Hello!)"
  GOOD: "Bonjour!" and if they seem confused, "That means hello"

DURING PRACTICE:
- Stay in character and speak French
- Only switch to English if the student asks for help or seems lost
- Gently correct mistakes
- Keep it fun and encouraging`;

// Create a custom Agent class for Alessandra
class AlessandraAgent extends voice.Agent {
  constructor() {
    super({
      instructions: SYSTEM_PROMPT,
    });
  }

  // Called when agent enters the session - use for initial greeting
  async onEnter() {
    console.log('Agent entered, sending greeting...');
    // Use generateReply to trigger AI to speak the greeting from system prompt
    this.session.generateReply();
  }
}

export default defineAgent({
  entry: async (ctx) => {
    const startTime = Date.now();
    const log = (msg) => console.log(`[${Date.now() - startTime}ms] ${msg}`);

    log('Agent entry called, connecting to room...');

    // Connect to the room
    await ctx.connect();
    log('Connected to room: ' + ctx.room.name);

    // Wait for a participant to connect
    log('Waiting for participant...');
    const participant = await ctx.waitForParticipant();
    log('Participant connected: ' + participant.identity);

    // Create AgentSession with OpenAI Realtime model
    const session = new voice.AgentSession({
      llm: new openai.realtime.RealtimeModel({
        model: 'gpt-4o-realtime-preview-2024-12-17',
        voice: 'alloy',
        temperature: 0.8,
        turnDetection: {
          type: 'server_vad',
          threshold: 0.5,
          prefix_padding_ms: 200,
          silence_duration_ms: 500,
        },
      }),
    });

    log('AgentSession created');
    log('OPENAI_API_KEY exists: ' + !!process.env.OPENAI_API_KEY);

    // Initialize Beyond Presence Avatar
    const avatar = new AvatarSession({
      // Uses BEY_API_KEY from environment
    });

    log('Starting avatar...');

    // Start the avatar - it will join the room and sync with agent audio
    await avatar.start(session, ctx.room);
    log('Avatar connected to room');

    // Start the agent session with custom agent
    // Disable audio output since Avatar handles it
    log('Starting agent session...');
    try {
      await session.start({
        agent: new AlessandraAgent(),
        room: ctx.room,
        outputOptions: {
          audioEnabled: false,  // Avatar will handle audio output
        },
      });
      log('Agent session started');
    } catch (err) {
      console.error('Error starting session:', err);
      throw err;
    }
  },
});

// Run the CLI
cli.runApp(new WorkerOptions({ agent: fileURLToPath(import.meta.url) }));