Personal Assistant (Jarvis) in Python


I thought it would be cool to create a personal assistant in Python. If you are into movies you may have heard of Jarvis, an A.I. based character in the Iron Man films. In this tutorial we will create a robot.

The features I want to have are:

For this tutorial you will need (Ubuntu) Linux, Python and a working microphone.

Related courses:

Video

This is what you’ll create:

Recognize spoken voice

Speech recognition can by done using the Python SpeechRecognition module. We make use of the Google Speech API because of it’s great quality.

Answer in spoken voice (Text To Speech)

Various APIs and programs are available for text to speech applications. Espeak and pyttsx work out of the box but sound very robotic. We decided to go with the Google Text To Speech API, gTTS.

sudo pip install gTTS

Using it is as simple as:

from gtts import gTTS
import os
tts = gTTS(text='Hello World', lang='en')
tts.save("hello.mp3")
os.system("mpg321 hello.mp3")

gtts

Complete program

The program below will answer spoken questions.

#!/usr/bin/env python3
# Requires PyAudio and PySpeech.
 
import speech_recognition as sr
from time import ctime
import time
import os
from gtts import gTTS
 
def speak(audioString):
    print(audioString)
    tts = gTTS(text=audioString, lang='en')
    tts.save("audio.mp3")
    os.system("mpg321 audio.mp3")
 
def recordAudio():
    # Record Audio
    r = sr.Recognizer()
    with sr.Microphone() as source:
        print("Say something!")
    audio = r.listen(source)
 
    # Speech recognition using Google Speech Recognition
    data = ""
    try:
        # Uses the default API key
        # To use another API key: `r.recognize_google(audio, key="GOOGLE_SPEECH_RECOGNITION_API_KEY")`
        data = r.recognize_google(audio)
        print("You said: " + data)
    except sr.UnknownValueError:
        print("Google Speech Recognition could not understand audio")
    except sr.RequestError as e:
        print("Could not request results from Google Speech Recognition service; {0}".format(e))
 
    return data
 
def jarvis(data):
    if "how are you" in data:
        speak("I am fine")
 
    if "what time is it" in data:
        speak(ctime())
 
    if "where is" in data:
        data = data.split(" ")
        location = data[2]
        speak("Hold on Frank, I will show you where " + location + " is.")
        os.system("chromium-browser https://www.google.nl/maps/place/" + location + "/&")
 
# initialization
time.sleep(2)
speak("Hi Frank, what can I do for you?")
while 1:
    data = recordAudio()
    jarvis(data)

Related posts:

 


String methods
Introduction to GUI

25 thoughts on “Personal Assistant (Jarvis) in Python

  1. Edward Principe - July 27, 2017

    Frank, I love the quality and execution of this program. I intend to build an interface to run some scientific equipment. I am not a programmer …. I generally hack my way through what I need to get the job done. I have written several basic programs to control the microscope.

    This is a Windows 8.1 system. Is that an issue??
    Installed the gTTS and SpeechRecognition. Having trouble getting PyAudio and PySpeech installed …. using python 3.3 and seems to need Visual C++ 10.0. Trying to work around that now. ….

    When I try to run your example code (short version), I get a string of errors, the end of which oddly seems tied to a URL related to ‘translate.google.com’…. if I interpret the error correctly.

      File "C:\Python33\lib\site-packages\requests\adapters.py", line 504, in send
        raise ConnectionError(e, request=request)
    requests.exceptions.ConnectionError: HTTPSConnectionPool(host='translate.google.com', port=443): Max retries exceeded with url: / (Caused by SSLError(SSLError(1, '[SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed (_ssl.c:547)'),))

    I know it is a mess …… any insights are appreciated!

    BTW: It is tts.save that generates the error.

    1. Frank - July 28, 2017

      Thanks Edward! Windows 8.1 should not be an issue, at the time I had tested it on Ubuntu.
      The gTTS module underneath uses the translate.google.com website, see inside the gtts source code. This website returns an audio file, which is played with any sound player (mpg321 as example).

      In this case I see a connection error, do you have a firewall? It may also be throttling (too many connections). If you have an offline environment, try ms sapi or espeak. The speech recognition part also needs internet connection though.