please provide a code in both front end and backend such that i can control deepgram transcriptions based on confidence level

To manage Deepgram transcriptions based on confidence levels, you can implement a custom transcriber with Vapi and evaluate each transcript's confidence before processing it. Here's a basic setup for the backend to handle audio with Deepgram and control output based on transcription confidence: - Backend Code (Node.js):

const { createClient, LiveTranscriptionEvents } = require("@deepgram/sdk");
const EventEmitter = require("events");
const express = require("express");
const http = require("http");

class TranscriptionService extends EventEmitter {
  constructor(apiKey) {
    super();
    this.deepgramClient = createClient(apiKey);
    this.deepgramLive = this.deepgramClient.listen.live({
      encoding: "linear16",
      channels: 2,
      sample_rate: 16000,
      models: "nova-2",
      smart_format: true,
      interim_results: true,
    });

    this.deepgramLive.on(LiveTranscriptionEvents.Transcript, (event) => {
      this.handleTranscript(event);
    });
  }

  handleTranscript(transcription) {
    const { channel_index, alternatives } = transcription.channel || {};
    if (!alternatives || alternatives[0].confidence < 0.85) return;  // Only process if confidence is above 85%
    const text = alternatives[0].transcript.trim();
    const channel = channel_index ? channel_index[0] : 0;
    console.log(`Channel ${channel}: ${text}`);
  }
}

// Express server setup
const app = express();
const server = http.createServer(app);

app.get("/", (req, res) => {
  res.send("Custom Transcriber Service is running");
});

const apiKey = process.env.DEEPGRAM_API_KEY;  // Your Deepgram API Key
const transcriptionService = new TranscriptionService(apiKey);

server.listen(3001, () => {
  console.log('Server running at http://localhost:3001');
});

- Frontend Integration: Use WebSocket to stream audio from the frontend to your backend server. Ensure audio is captured in linear16 format, with a sample rate of 16000 Hz. Ensure you have the following dependencies installed: @deepgram/sdk, express, and ws. For detailed implementation, refer to the Custom Transcriber documentation which explains setting up a server using Vapi with Deepgram. Feel free to adjust the confidence threshold or logging level to fit your specific requirements.