Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/Harsha200105/DesktopAssistant/llms.txt

Use this file to discover all available pages before exploring further.

Jarvis listens for your voice through the microphone, converts speech to text using Google’s Speech Recognition API, matches the transcribed text against known command phrases, and executes the corresponding action. Every command follows the same loop: speak a phrase → Jarvis recognizes it → the action runs → Jarvis prompts “Next Command Sir!” so you can issue another.

​
How the assistant listens

When you speak, Jarvis uses the speech_recognition library to capture audio and send it to Google’s Speech API. The recognizer is tuned with a short pause threshold to feel responsive: 
def take_command():
    r = sr.Recognizer()
    with sr.Microphone() as source:
        print("Listening....")
        r.pause_threshold = 0.5
        r.energy_threshold = int(config['DEFAULT']['energy_threshold'])
        audio = r.listen(source)

    query = " "
    try:
        print("Recognizing....")
        query = r.recognize_google(audio, language="en-in")
        print("user said: " + query)

    except sr.UnknownValueError:
        speak("Sorry Could You please try again")

    except Exception as e:
        print(e)

    return query
ParameterValueNotes
pause_threshold0.5 secondsSilence needed before Jarvis stops recording
energy_threshold300 (default)Microphone sensitivity — raise if Jarvis misses commands; lower if it never triggers
languageen-inEnglish (India) locale passed to r.recognize_google()
The energy_threshold value is read from config.ini so you can tune it without editing source code.
If Jarvis frequently says “Sorry, could you please try again?” without you even speaking, your microphone’s background noise level exceeds energy_threshold. Increase the value in config.ini under [DEFAULT] → energy_threshold.

​
Command categories

Web Navigation

Open websites by name and search the web with configurable search engines including Google, Bing, DuckDuckGo, and YouTube.

Information

Get Wikipedia summaries, ask for the current date or time, and receive friendly conversational responses.

Music

Play, pause, resume, and stop local MP3 music using pygame mixer — all by voice.

Email

Dictate and send emails via SMTP by speaking the recipient address and message body.

System Commands

Greet Jarvis, change voice gender, speech rate, volume, and exit the assistant gracefully.

​
Quick reference

The table below lists every recognized phrase and what it does. Phrases are matched case-insensitively on the lowercased transcript.
PhraseAction
helloGreet the user — Jarvis replies “Hello Sir”
what's upRandom friendly response from a preset list
open [website]Open a built-in website (Google, YouTube, Wikipedia, Amazon, GitHub) or offer to search for an unknown site
search for [query]Run a web search using the configured search engine
wikipedia [topic]Fetch and speak a 2-sentence Wikipedia summary of the topic
timeSpeak the current time in 12-hour format — Ubuntu (Jarvis2.py) only
dateSpeak the current date — Ubuntu (Jarvis2.py) only
mailStart the email flow — Jarvis asks for recipient and message body
play musicLoad and play a random MP3 from your configured music folder
pause musicPause the currently playing track
unpauseResume a paused track
stop musicStop music playback entirely
change voice to maleSwitch TTS voice to the first system voice (male)
change voice to femaleSwitch TTS voice to the second system voice (female)
change rate to [number]Set the TTS speech rate (e.g. “change rate to 150”)
change volume to [number]Set the TTS volume 0–100 (e.g. “change volume to 80”)
nothing / abort / stopSay goodbye and exit the program
byeSay goodbye and exit the program

​
The “Next Command Sir” prompt

After every command executes, Jarvis always speaks “Next Command! Sir!” — this is the signal that the assistant is ready and listening again. The prompt is unconditional: it fires at the end of execute_the_command_said_by_user() regardless of which branch ran. 
    # ... command logic above ...

    speak("Next Command! Sir!")
In debug mode (debug = True in config.ini), Jarvis reads commands from input() at the terminal instead of the microphone. This is useful for testing without a microphone or in noisy environments.

Build docs developers (and LLMs) love

Get started for freeTalk to us