### The Ultimate Cross-Platform Personal AI Assistant — By FatihMakes
> 📺 **[Watch the full setup video on YouTube](https://www.youtube.com/@FatihMakes)**
A real-time voice AI that can hear, see, understand, and control your computer — on any OS. Supports Windows, macOS, and Linux. Built on the Gemini Live API for native audio streaming, delivering zero subscriptions and total digital autonomy.
MARK XLIX deepens the personal assistant foundation. Rather than adding more tools, this build focused on making the assistant truly *yours*: it starts with your computer, learns your name, and pays attention to what you're doing. The goal before the plugin era begins is a core that feels alive — not just reactive.
The assistant now registers itself with the operating system's startup system. One click in the UI toggles it on or off. On Windows, it writes to the registry using `pythonw.exe` so no console window ever appears. On macOS it installs a LaunchAgent plist; on Linux a `.desktop` autostart entry. The button reflects the current state every time the app launches.
The assistant is no longer locked to the name "JARVIS". Click `⚙ CUSTOMISE ASSISTANT` in the right panel to change:
- **Assistant name** — displayed everywhere in the UI (title bar, header, HUD, log, footer) and injected into the Gemini system prompt so the AI knows its own name
- **Your name** — how the assistant addresses you. Leave blank for the default language-aware addressing (`sir` / `efendim`), or set your actual name for a more personal feel
Copy any text of 10 or more characters and a floating panel appears at the bottom of the window. Four quick actions — **TRANSLATE**, **SUMMARISE**, **EXPLAIN**, **FIX** — send the copied content directly to the assistant with one click. The panel auto-dismisses after 8 seconds. This turns the clipboard into a silent command channel for anything on your screen.
The morning briefing can now be turned on or off with one click from the settings drawer (`⚙` → `☀ MORNING BRIEF: ON/OFF`). Users who don't want a startup briefing can disable it permanently; the setting survives restarts. The briefing itself was also re-engineered: news is now pre-fetched in a background thread the moment the session starts, running in parallel while the greeting plays. By the time the greeting finishes, the results are already ready — no extra Gemini tool-call round-trip needed. Briefing delivery is noticeably faster as a result.
> ⚠️ **Installation Note:** Some OS-specific dependencies are not bundled in `requirements.txt` to keep the repo lightweight. If you hit a `ModuleNotFoundError`, install the missing package with `pip install <module_name>`.
---
## 📋 Requirements
| Requirement | Details |
| --- | --- |
| **OS** | Windows 10/11, macOS, or Linux |
| **Python** | 3.11 or 3.12 |
| **Microphone** | Required for voice interaction |
| **API Key** | Free Gemini API key (`config/api_keys.json`) |