Jarvis is an AI desktop assistant capable of listening, understanding, remembering, and responding to voice commands in real time. It combines speech recognition, an AI planner, long-term memory, and voice responses into a natural conversational experience.

Real-time voice interaction

AI-powered conversation

Long-term memory

Natural text-to-speech responses

Date and time queries

Web search

Open Windows applications

Basic safety filtering

Multi-threaded architecture for smooth interaction

Python

Groq API

Whisper (Speech-to-Text)

pyttsx3

sounddevice

faster-whisper (experimental)

JSON Memory Store

Jarvis can currently:

Answer general questions

Remember user information

Recall saved memories

Open Windows applications

Search the web

Tell the date and time

Hold natural conversations

Better speech recognition

Smarter application detection

Improved conversational memory

File management

Browser automation

Email support

Calendar integration

Vision (screen understanding)

Wake word detection

Local AI fallback

Build an AI operating system assistant comparable to the fictional Jarvis from Iron Man—capable of understanding context, assisting with everyday work, and interacting naturally with a computer.

Create a .env file:

Run:

Early Beta

The project is functional but under active development. Features, APIs, and architecture may change frequently.

Suggestions, issues, and pull requests are welcome.

MIT License

David Adeleke

Building Jarvis in public.

Follow the journey on X: https://x.com/adeleke8498