Where Ancient Wisdom Meets Modern AI
प्राचीन ज्ञान • आधुनिक बुद्धि
A fully offline, voice-controlled personal AI assistant with a warm Hinglish personality, built to guide, teach, and take care of you — running entirely on your own machine.
SYTA — warm, offline, always ready to help
Building a personalized 9-month learning roadmap — fully offline
---SYTA (featuring the assistant Sita) is a personal AI companion that lives on your computer. Unlike cloud assistants, SYTA runs 100% offline after setup — your conversations never leave your machine.
Sita isn't just a command bot. She's designed to feel like a caring guide: she remembers your goals, teaches you new skills step by step, checks in on you during the day, and helps you control your computer by voice — all in a natural Hinglish (Hindi + English) personality.
- 🗣️ Voice control — Speak to Sita, she speaks back. Say "Hey Sita" anytime.
- 🧠 Local AI brain — Powered by Llama 3 via Ollama. No internet needed, fully private.
- 💻 System control — Open apps, adjust brightness, check battery/storage, find files, lock screen, take screenshots — all by voice.
- 📅 Meeting manager — Add meetings, get reminders 30 minutes before.
- 📚 Learning coach — Ask Sita to teach you a skill; she builds a step-by-step roadmap.
- 💛 Proactive care — Morning briefings, evening check-ins, health reminders.
- 🎭 Animated face — A friendly on-screen avatar that blinks, looks around, and talks.
- 🧵 Long-term memory — Sita remembers your goals and past conversations across restarts.
| Component | Technology |
|---|---|
| Language | Python 3.11 |
| AI Model | Llama 3 (via Ollama) |
| Speech-to-Text | faster-whisper |
| Text-to-Speech | pyttsx3 |
| GUI | Tkinter |
| System control | psutil, pyautogui, screen-brightness-control |
- Python 3.11 — Download here (check "Add Python to PATH")
- Ollama — Download here
# 1. Clone this repository
git clone https://github.com/YOUR_USERNAME/syta.git
cd syta
# 2. Install Python dependencies
pip install -r requirements.txt
# 3. Download the AI model (one-time, ~4.7 GB)
ollama pull llama3
# 4. Run SYTA
python sita_core.pyNote: On first run, Ollama needs to be running in the background. Start it with
ollama serveif it isn't already.
Once SYTA is running, you can type or click the MIC button to talk to Sita. Try:
| You say | Sita does |
|---|---|
| "battery check karo" | Reports battery status |
| "brightness 70 karo" | Sets screen brightness |
| "find file resume" | Searches and opens the file |
| "open chrome" | Launches Chrome |
| "I want to learn Python" | Builds a learning roadmap |
| "Hey Sita" | Wakes her up to listen |
- Streaming AI responses (show text as it generates)
- Better Hinglish voice (edge-tts neural voices)
- Confirmation gates for destructive actions
- System tray mode with global hotkey
- Timers and alarms
- Modular refactor (separate tts, stt, actions, memory modules)
This is a personal project under active development. Suggestions, bug reports, and pull requests are welcome! Open an issue to start a discussion.
This project is released under the MIT License — see LICENSE for details.
Model note: Llama 3 is used under Meta's community license. faster-whisper and other dependencies carry their own licenses — review them before commercial use.
Built with the help of Claude (Anthropic). Inspired by the vision of blending India's ancient wisdom traditions with modern AI technology.
Made with 💛 by Abhinav
SYTA — Always here. Always yours.