[IoT] MirAI – Intelligent & Smart Mirror Assistant

IoT>Project
3.8k words

Overview

MirAI is an intelligent connected mirror that enhances daily life through voice interactions, real-time information, and smart home integration. Using Azure Cognitive Services and OpenAI, MirAI provides a seamless user experience with natural language processing, weather updates, calendar synchronization, and much more.

My Role

  • Implemented voice recognition and speech synthesis using Azure Speech Services.
  • Integrated OpenAI for intelligent conversation handling.
  • Designed and built real-time weather synchronization.
  • Developed a voice-activated user interface for smooth interaction.

Tools & Technologies

✅ Azure Cognitive Services (Speech-to-Text, Text-to-Speech)
✅ OpenAI API (Chat-based interactions)
✅ Python (Core development language)
✅ SimpleAudio (Audio feedback system)
✅ Requests (Fetching weather data from OpenWeather API)
✅ Dotenv (Managing environment variables)

Code Highlights

Speech Recognition with Azure Speech Services

speech_recognition.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
  
import azure.cognitiveservices.speech as speechsdk  

def Start_recording(output_folder):  
    speech_config = speechsdk.SpeechConfig(
        subscription=settings['speechKey'], region=settings['region'], 
        speech_recognition_language=settings['language']
    )  

    audio_config = speechsdk.audio.AudioConfig(use_default_microphone=True)  
    speech_recognizer = speechsdk.SpeechRecognizer(speech_config=speech_config, audio_config=audio_config)  

    print("Listening...")  
    result = speech_recognizer.recognize_once()  

    return result.text if result.reason == speechsdk.ResultReason.RecognizedSpeech else "Speech not recognized."

OpenAI Chat Integration

openai_integration.py
1
2
3
4
5
6
7
8
9
10
11
12
13
  
import openai  

openai.api_key = "YOUR_OPENAI_KEY"  

def complete_openai(name, prompt, token=100):  
    response = openai.completions.create(  
        model="gpt-3.5-turbo-instruct",  
        prompt=f"You are MirAI, a smart mirror assistant. Address the user by name: {name}. Here is the input: {prompt}",  
        temperature=0.7,  
        max_tokens=token  
    )  
    return response.choices[0].text.strip()

Fetching Weather Data

weather_fetch.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
  
import requests  
import os  

API_KEY = os.getenv("OPENWEATHER_KEY")  

def get_weather_forecast(city):  
    url = f"http://api.openweathermap.org/data/2.5/weather?q={city}&units=imperial&appid={API_KEY}"  
    response = requests.get(url).json()  

    if response.get("cod") != 200:  
        return "City not found."  

    weather = response["weather"][0]["main"]  
    temp_fahrenheit = response["main"]["temp"]  
    temp_celsius = round((temp_fahrenheit - 32) * 5/9)  

    return f"The weather in {city} is {weather}, with a temperature of {temp_celsius}°C."

Processing User Requests

ask_request.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
  
def AskRequest(name):  
    user_input = Start_recording(output_folder=output_folder)  
    conversation.append(user_input)  
    print(user_input)  

    result = ""  

    user_input_lower = user_input.lower()  
    if 'current temperature' in user_input_lower:  
        prompt = f"Based on this input: '{user_input}', extract only the city name, without quotes or formatting."  
        city_result = complete_openai(name, prompt=prompt, token=3000)  

        prompt = f"Based on this weather data: {get_weather_forecast(city_result)}, generate a short assistant-like response."  
        result = complete_openai(name, prompt=prompt, token=3000)  

    else:  
        result = complete_openai(name, prompt=user_input, token=3000)  

    conversation.append(result)  
    speak(result, output_folder=output_folder)

Playing a Confirmation Sound

play_sound.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
  
import simpleaudio as sa  
import numpy as np  

def play_sound():  
    frequency = 440  
    duration = 0.1  
    sample_rate = 44100  
    amplitude = 16000  

    waveform = np.sin(2 * np.pi * np.arange(sample_rate * duration) * frequency / sample_rate)  
    waveform = (waveform * amplitude).astype(np.int16)  

    audio = sa.play_buffer(waveform, 1, 2, sample_rate)  
    audio.wait_done()

MirAI transforms a regular mirror into an AI-powered assistant, making daily routines smoother and more interactive.