A developer has built Jarvis, a personal AI assistant written entirely in Python as a learning project. The goal was to understand real-world assistant architecture, from voice input to intent handling. It features SpeechRecognition, pyttsx3 for text-to-speech, and a Tkinter GUI for text input.

The system uses a rule-based intent detection with an optional OpenAI or Gemini LLM fallback. A modular structure with single-responsibility files like `voice.py` and `automation.py` makes the code extensible. Current capabilities include wake-word activation, continuous conversation, and basic Windows-focused system automation.

Key limitations are the lack of persistent memory and limited built-in skills. The developer plans to add a plugin-based skill system, better intent classification, and offline command support. Future work includes packaging as a Windows executable and exploring SQLite for memory. The open-source project is on GitHub for community feedback.