🎙️ Thai Real-time Speech-to-Text (ASR) 🆕

🎙️ ระบบแปลงเสียงพูดภาษาไทยเป็นข้อความแบบเรียลไทม์ (Real-time)

Welcome to Thai Real-time ASR - our WebSocket-based real-time Automatic Speech Recognition service. This service converts audio stream data into text stream data in real-time, supporting both file streaming and direct microphone capture.

Getting Started

1. Prerequisites

   • An API key from iApp Technology
   • Audio stream in supported format
   • WebSocket client implementation
   • Supported format: PCM audio
   • No specific length limitation

2. Quick Start

   • Real-time processing
   • Stream from file or microphone
   • Support for Thai language

3. Key Features

   • Real-time text streaming
   • WebSocket-based communication
   • Low latency processing
   • Flexible integration options
   • Continuous speech recognition

4. Security & Compliance

   • GDPR and PDPA compliant
   • Secure WebSocket (WSS) support
   • No data retention after processing

How to get API Key?

Please visit API Portal to view your existing API key or request a new one.

Example

Sample Request:

• Without encryption : ws://api.iapp.co.th/asrx
• With encryption : wss://api.iapp.co.th/asrx

Sample Response:

{
    "type": "realtime",
    "state": "Sentence",
    "text": "สวัสดีครับ"
}

Features & Capabilities

Core Features

• Real-time speech recognition
• WebSocket streaming support
• File and microphone input
• State-based response system
• Continuous recognition

Supported Fields

• Audio stream data
• PCM audio format
• WebSocket protocols
• Real-time text output
• State indicators

API Reference

• Endpoints:
  • Without encryption : ws://api.iapp.co.th/asrx
  • With encryption : wss://api.iapp.co.th/asrx
• Required Headers:
  • apikey: Your API key

Audio Format Requirements

• Sample Rate: 16,000 Hz
• Channel: Mono
• Bit-depth: 16 bit
• Audio-encode: PCM

Code Examples

Python (Microphone Streaming)

import pyaudio
import websockets
import asyncio

async def stream_microphone():
    url = "wss://api.iapp.co.th/asrx"
    audio = pyaudio.PyAudio()
    stream = audio.open(format=pyaudio.paInt16,
                        channels=1,
                        rate=16000,
                        input=True,
                        frames_per_buffer=1024)

    async with websockets.connect(url, extra_headers={"apikey": "YOUR_API_KEY"}) as ws:
        while True:
            data = stream.read(1024)
            await ws.send(data)
            await asyncio.sleep(0.05)

Python (File Streaming)

from websockets.client import WebSocketClientProtocol
import websockets
import asyncio
import json

async def send_audio_data(ws_client: WebSocketClientProtocol):
    frame_size = 1024
    with open("audio.pcm", "rb") as file:
        while True:
            buff = file.read(frame_size)
            if buff:
                await ws_client.send(buff)
            else:
                break
            await asyncio.sleep(0.02)

async def client_test():
    url = "wss://api.iapp.co.th/asrx"
    async with websockets.connect(url, extra_headers={"apikey": "YOUR_API_KEY"}) as ws:
        await send_audio_data(ws)

Javascript

const WebSocket = require('ws'); // Ensure WebSocket is installed (npm install ws)
const fs = require('fs');

async function sendAudioData(ws) {
    const frameSize = 1024; // Define the frame size for audio chunks
    const fileStream = fs.createReadStream('audio.pcm', { highWaterMark: frameSize });

    // Read and send audio data frame by frame
    for await (const chunk of fileStream) {
        if (chunk) {
            ws.send(chunk); // Send the audio data
        }
        await new Promise(resolve => setTimeout(resolve, 20)); // Wait for 20 ms
    }

    // Close the WebSocket connection after sending all data
    ws.close();
}

function clientTest() {
    const url = "wss://api.iapp.co.th/asrx";
    const apiKey = "YOUR_API_KEY";

    // Connect to the WebSocket server
    const ws = new WebSocket(url, {
        headers: {
            apikey: apiKey
        }
    });

    ws.on('open', async () => {
        console.log('WebSocket connection opened');
        await sendAudioData(ws);
    });

    ws.on('message', (message) => {
        console.log('Received:', message.toString());
    });

    ws.on('error', (error) => {
        console.error('WebSocket error:', error);
    });

    ws.on('close', (code, reason) => {
        console.log('WebSocket closed:', code, reason);
    });
}

// Run the client
clientTest();

PHP

<?php
$apiKey = "YOUR_API_KEY";
$url = "wss://api.iapp.co.th/asrx";

$ws = new WebSocket($url, ['headers' => ["apikey: $apiKey"]]);

function sendAudioData($ws) {
    $frameSize = 1024; // Frame size in bytes
    $file = fopen("audio.pcm", "rb");

    while (!feof($file)) {
        $data = fread($file, $frameSize);
        $ws->send($data);
        usleep(20000); // 20ms delay
    }

    fclose($file);
    $ws->close();
}

$ws->on("open", function($ws) {
    sendAudioData($ws);
});

$ws->on("message", function($message) {
    echo "Received: $message\n";
});

$ws->on("error", function($error) {
    echo "Error: $error\n";
});

$ws->on("close", function() {
    echo "WebSocket closed\n";
});

Swift

import Foundation

class WebSocketClient: NSObject, URLSessionWebSocketDelegate {
    let apiKey = "YOUR_API_KEY"
    let url = URL(string: "wss://api.iapp.co.th/asrx")!
    var webSocket: URLSessionWebSocketTask?

    func connect() {
        let session = URLSession(configuration: .default, delegate: self, delegateQueue: OperationQueue())
        var request = URLRequest(url: url)
        request.addValue(apiKey, forHTTPHeaderField: "apikey")

        webSocket = session.webSocketTask(with: request)
        webSocket?.resume()

        sendAudioData()
    }

    func sendAudioData() {
        guard let webSocket = webSocket else { return }
        let fileURL = URL(fileURLWithPath: "audio.pcm")

        guard let fileHandle = try? FileHandle(forReadingFrom: fileURL) else {
            print("Could not open audio file.")
            return
        }

        DispatchQueue.global().async {
            while true {
                let data = fileHandle.readData(ofLength: 1024)
                if data.isEmpty { break }

                let message = URLSessionWebSocketTask.Message.data(data)
                webSocket.send(message) { error in
                    if let error = error {
                        print("Error sending data: \(error)")
                        return
                    }
                }

                Thread.sleep(forTimeInterval: 0.02)
            }

            fileHandle.closeFile()
            webSocket.cancel()
        }
    }

    func urlSession(_ session: URLSession, webSocketTask: URLSessionWebSocketTask, didCloseWith closeCode: URLSessionWebSocketTask.CloseCode, reason: Data?) {
        print("WebSocket closed")
    }
}

let client = WebSocketClient()
client.connect()

Kotlin

import okhttp3.OkHttpClient
import okhttp3.Request
import okhttp3.WebSocket
import okhttp3.WebSocketListener
import okio.ByteString
import java.io.File
import kotlinx.coroutines.*

class WebSocketClient : WebSocketListener() {
    private val apiKey = "YOUR_API_KEY"
    private val url = "wss://api.iapp.co.th/asrx"

    fun connect() {
        val client = OkHttpClient()
        val request = Request.Builder()
            .url(url)
            .addHeader("apikey", apiKey)
            .build()

        val webSocket = client.newWebSocket(request, this)

        CoroutineScope(Dispatchers.IO).launch {
            sendAudioData(webSocket)
        }
    }

    private suspend fun sendAudioData(webSocket: WebSocket) {
        val frameSize = 1024
        val file = File("audio.pcm")
        val byteArray = ByteArray(frameSize)

        file.inputStream().use { inputStream ->
            while (inputStream.read(byteArray) != -1) {
                webSocket.send(ByteString.of(byteArray, 0, frameSize))
                delay(20) // 20ms delay
            }
        }
        webSocket.close(1000, "Finished sending data")
    }

    override fun onMessage(webSocket: WebSocket, text: String) {
        println("Received: $text")
    }

    override fun onFailure(webSocket: WebSocket, t: Throwable, response: okhttp3.Response?) {
        println("Error: ${t.message}")
    }

    override fun onClosing(webSocket: WebSocket, code: Int, reason: String) {
        println("WebSocket closing: $code / $reason")
        webSocket.close(code, reason)
    }
}

fun main() {
    val client = WebSocketClient()
    client.connect()
}

Java

import okhttp3.*;
import okio.ByteString;

import java.io.File;
import java.io.FileInputStream;
import java.io.IOException;

public class WebSocketClient {
    private static final String URL = "wss://api.iapp.co.th/asrx";
    private static final String API_KEY = "YOUR_API_KEY";

    public static void main(String[] args) {
        OkHttpClient client = new OkHttpClient();
        Request request = new Request.Builder()
                .url(URL)
                .addHeader("apikey", API_KEY)
                .build();

        WebSocket webSocket = client.newWebSocket(request, new WebSocketListener() {
            @Override
            public void onOpen(WebSocket webSocket, Response response) {
                System.out.println("WebSocket opened");
                new Thread(() -> sendAudioData(webSocket)).start();
            }

            @Override
            public void onMessage(WebSocket webSocket, String text) {
                System.out.println("Received: " + text);
            }

            @Override
            public void onFailure(WebSocket webSocket, Throwable t, Response response) {
                System.out.println("Error: " + t.getMessage());
            }

            @Override
            public void onClosing(WebSocket webSocket, int code, String reason) {
                System.out.println("WebSocket closing: " + code + " / " + reason);
                webSocket.close(code, reason);
            }
        });
    }

    private static void sendAudioData(WebSocket webSocket) {
        int frameSize = 1024;
        File file = new File("audio.pcm");
        try (FileInputStream fis = new FileInputStream(file)) {
            byte[] buffer = new byte[frameSize];
            int bytesRead;
            while ((bytesRead = fis.read(buffer)) != -1) {
                webSocket.send(ByteString.of(buffer, 0, bytesRead));
                Thread.sleep(20); // 20ms delay
            }
            webSocket.close(1000, "Finished sending data");
        } catch (Exception e) {
            e.printStackTrace();
        }
    }
}

Dart

import 'dart:async';
import 'dart:convert';
import 'dart:io';

void main() async {
  final url = "wss://api.iapp.co.th/asrx";
  final apiKey = "YOUR_API_KEY";

  final socket = await WebSocket.connect(url, headers: {'apikey': apiKey});
  print('WebSocket connected');

  sendAudioData(socket);
  socket.listen(
    (data) => print('Received: $data'),
    onError: (error) => print('Error: $error'),
    onDone: () => print('WebSocket closed'),
  );
}

void sendAudioData(WebSocket socket) async {
  final file = File('audio.pcm');
  final frameSize = 1024;
  final raf = file.openRead();

  await for (final chunk in raf) {
    socket.add(chunk);
    await Future.delayed(Duration(milliseconds: 20)); // 20ms delay
  }

  socket.close();
}

Limitations and Best Practices

Limitations

• Thai language support only
• Specific audio format requirements
• WebSocket connection required
• Network dependency

Best Practices

• Maintain stable network connection
• Use appropriate frame size
• Handle connection errors gracefully
• Monitor WebSocket state
• Process response states correctly

Accuracy & Performance

Overall Accuracy

• Real-time transcription quality
• Continuous speech recognition
• State-based accuracy tracking

Processing Speed

• Real-time processing
• Low latency response
• Network-dependent performance

Factors Affecting Accuracy

• Audio quality
• Network stability
• Background noise
• Speaking clarity
• Microphone quality

History

Version 1.0 (2023)

• Initial release
• WebSocket streaming support
• Real-time processing capability
• File and microphone input support
• State-based response system

Pricing

AI API Service Name	Endpoint	IC Per Seconds	On-Premise
Thai Speech To Text (ASR)	iapp-asr-v3-en	1 IC/60 Seconds	Contact
	iapp-asr-v3-th-en	1 IC/60 Seconds