Building a Python Voice Chatbot with Baidu AI Speech Recognition and Qingyunke

In recent days I needed a small Python program that enables a human to have a voice conversation with an intelligent robot using a smart dialogue interface. The overall workflow consists of four steps: capture voice input, convert the audio to text, call an AI dialogue API to obtain a textual response, and finally convert that response back to speech.

Required environment : install the following Python packages:

pip install pyaudio – for recording and generating WAV files

pip install baidu-aip – Baidu AI SDK for speech‑to‑text

pip install pyttsx3 – text‑to‑speech synthesis

Recording voice and saving as WAV :

import time
import wave
from pyaudio import PyAudio, paInt16

framerate = 16000  # sampling rate
num_samples = 2000  # frames per buffer
channels = 1        # mono
sampwidth = 2      # 2 bytes per sample
FILEPATH = '../voices/myvoices.wav'  # ensure directory exists

class Speak():
    def save_wave_file(self, filepath, data):
        wf = wave.open(filepath, 'wb')
        wf.setnchannels(channels)
        wf.setsampwidth(sampwidth)
        wf.setframerate(framerate)
        wf.writeframes(b''.join(data))
        wf.close()

    def my_record(self):
        pa = PyAudio()
        stream = pa.open(format=paInt16, channels=channels, rate=framerate, input=True, frames_per_buffer=num_samples)
        my_buf = []
        t = time.time()
        print('正在讲话...')
        while time.time() < t + 5:  # record for 5 seconds
            string_audio_data = stream.read(num_samples)
            my_buf.append(string_audio_data)
        print('讲话结束')
        self.save_wave_file(FILEPATH, my_buf)
        stream.close()

Speech‑to‑text with Baidu AI (you need APP_ID, API_KEY and SECRET_KEY from the Baidu AI console):

from aip import AipSpeech

APP_ID = '25990397'
API_KEY = 'iS91n0uEOujkMIlsOTLxiVOc'
SECRET_KEY = ''  # fill with your secret key

client = AipSpeech(APP_ID, API_KEY, SECRET_KEY)

class ReadWav():
    def get_file_content(self, filePath):
        with open(filePath, 'rb') as fp:
            return fp.read()

    def predict(self):
        return client.asr(self.get_file_content('../voices/myvoices.wav'), 'wav', 16000, {'dev_pid': 1537})

readWav = ReadWav()
print(readWav.predict())  # example output contains the recognized text

Chatbot interaction using Qingyunke (free GET‑based API):

def talkWithRobot(msg):
    url = 'http://api.qingyunke.com/api.php?key=free&appid=0&msg={}'.format(urllib.parse.quote(msg))
    html = requests.get(url)
    return html.json()["content"]

print(talkWithRobot("你好呀!"))  # → 哟~ 都好都好

Text‑to‑speech synthesis with pyttsx3 :

import pyttsx3

class RobotSay():
    def __init__(self):
        self.engine = pyttsx3.init()
        self.rate = self.engine.getProperty('rate')
        self.engine.setProperty('rate', self.rate - 50)

    def say(self, msg):
        self.engine.say(msg)
        self.engine.runAndWait()

robotSay = RobotSay()
robotSay.say("你好呀")  # plays the spoken text

Putting everything together into an interactive loop:

robotSay = RobotSay()
speak = Speak()
readTalk = ReadWav()

while True:
    speak.my_record()                               # record voice
    text = readTalk.predict()['result'][0]            # speech‑to‑text
    print("本人说:", text)
    response_dialogue = talkWithRobot(text)          # chatbot reply
    print("青云客说:", response_dialogue)
    robotSay.say(response_dialogue)                  # speak the reply

The program continuously records the user’s speech, converts it to text via Baidu AI, obtains a chatbot reply from Qingyunke, and reads the reply aloud. Sample console output demonstrates a back‑and‑forth conversation.

Future work includes adding a graphical user interface and extending functionality.