Monika-AI is a high-performance, personality-driven interactive companion. Leveraging Gemini 1.5 Flash and advanced browser APIs, Monika doesn't just chatβshe sees, hears, and follows you across your desktop with a custom Picture-in-Picture interface.
- ποΈ Vision Engine: Monika can now see through your webcam. Show her objects, code, or your surroundings, and she will react in real-time.
- πΌοΈ Floating Window (PiP): Multitask with ease. Use the Document Picture-in-Picture API to pop Monika into a floating, always-on-top window while you code or game.
- π Cyber-Sakura UI: A gorgeous "Glassmorphism" interface featuring:
- 3D Mouse Parallax: The UI panels tilt and react to your cursor movement.
- Dynamic Mood Engine: Backgrounds and glows shift colors based on Monika's emotions (
[HAPPY],[LOVING],[THINKING]). - Scanline Effects: A sci-fi HUD overlay for the vision feed.
- ποΈ Seamless Voice Loop: Zero-latency speech synthesis combined with hands-free
SpeechRecognition. - πΎ Persistent Memory: MongoDB integration ensures she remembers your name, projects (like your Pothole Portal or CandyRobot), and past conversations.
- Document PiP API: Advanced window detachment for "Always-on-Top" functionality.
- CSS3 Glassmorphism: Deep blurs, saturation filters, and
cubic-beziertransitions. - Web Speech API: Native browser processing to keep server CPU usage low.
- Gemini 1.5 Flash: Optimized for high-speed multimodal (Text + Image) reasoning.
- Buffered Image Processing: Secure Base64 handling for vision frames.
- Mongoose ODM: Structured conversation logging and user profiling.
Monika-AI/
βββ backend/
β βββ server.js # Express, Gemini Vision Logic & MongoDB
β βββ package.json # Node dependencies
βββ public/
β βββ index.html # Multimodal UI & PiP entry point
β βββ style.css # 3D effects, Glassmorphism & Mood themes
β βββ script.js # Vision capture, PiP Logic & Voice Loop
-
Clone & Install
git clone [https://github.com/tagadearpit/Monika-AI.git](https://github.com/tagadearpit/Monika-AI.git) cd Monika-AI/backend npm install -
Environment Setup Create a
.envfile in/backend:PORT=3000 GEMINI_API_KEY=your_key_here MONGO_URI=your_mongodb_atlas_uri
-
Run
npm start
Recommended: Access via Google Chrome for full Vision and PiP support.
Monika's environment reacts to her internal state:
- [HAPPY]: Soft violet-blue gradients.
- [LOVING]: Pulsing pink "Heart-Glow" animations.
- [THINKING]: Deep space-blue with increased scanline intensity.
Have ideas for a more "glitchy" aesthetic or better vision prompts? PRs are welcome!
Developed with β€οΈ by Arpit Tagade
### π‘ What changed in this version:
1. **Updated Badges:** Switched to `for-the-badge` style for a more "premium" feel.
2. **Vision & PiP Focus:** Put your two coolest technical features (Vision and Floating Window) at the very top.
3. **Modernized Tech Stack:** Mentioned the specific APIs (Document PiP, Glassmorphism) that recruiters love to see.
4. **Hardware Context:** Subtle references to her remembering your specific projects, which shows off her "Long-Term Memory."