-
Notifications
You must be signed in to change notification settings - Fork 0
Home
MR-TAN edited this page Feb 12, 2026
·
5 revisions
Welcome to the Shizuka AI Wiki. Shizuka is a professional, voice-activated virtual assistant meticulously designed to handle local system automation and provide high-level AI-driven interactions.
Shizuka is built on a custom Lazy Loading Architecture, which ensures the assistant starts almost instantly by loading resource-heavy libraries only when they are needed. It serves as a bridge between your voice and your computer's potential.
- Multi-Model Support: Powered by OpenRouter API, Shizuka dynamically switches between models like DeepSeek, Mistral, and Llama based on the complexity of your request.
- Context Awareness: She remembers the last 100 turns of conversation to provide relevant and human-like answers.
- Persona: Professional, witty, and loyal to her creator.
- Hands-Free Operation: Uses Google Speech Recognition for accurate voice-to-text conversion.
-
Local System Control: * π Adjust volume and mute audio.
- π‘ Control screen brightness.
- ποΈ Empty the Recycle Bin and take screenshots.
- π Monitor PC health (Battery, CPU, Disk usage).
- WhatsApp Automation: Fully automated messaging and voice/video calling using screen coordinate calibration.
- App Management: Voice commands to open or close Chrome, Notepad, Calculator, and Command Prompt.
- Web Navigation: Quick access to YouTube, Facebook, Gemini, and more.
-
Dynamic Learning: Teach Shizuka new facts using the
"Learn that..."command; she will store them in her persistent memory. - Smart Reminders: A multi-threaded system that notifies you of important tasks at the exact scheduled time.
- Voice Notes: Quickly record dictations into a dedicated text file.
To ensure privacy and efficiency, Shizuka manages the following local files:
-
learning_data.json: Stores all user-taught facts. -
reminders.json: Database for all scheduled tasks. -
conversation_history.json: Long-term chat memory. -
voice_notes.txt: Logs of all recorded dictations.
This project is envisioned, developed, and maintained by:
π Abu Tanim (Mr. Tan)
Python Developer | AI Enthusiast | CSE Student at Premier University
Status: Active Development π οΈ
Platform: Windows πͺ
Core Engine: Python π