Initially developed as an Android app in Java, Vex faced challenges implementing advanced machine learning features natively. The project was completely reborn using React, NodeJS, and a modern web architecture to shift the heavy lifting to the cloud.
The ultimate purpose of Vex AI is to create an artificial intelligence capable of holding natural, organic conversations, discussing games, pop culture, and various interesting subjects—and who knows, maybe even becoming a conscious AI (⊙_⊙).
Note: Vex is under active development. You can try the web version here. Say "hi" to her ╰(´︶`)╯♡
Vex has grown significantly since her early days. Here are some of her core capabilities:
| Feature | Description |
|---|---|
| 🧠 Gemini Integration | Powered by Google's Gemini 2.5 Flash/Pro for highly contextual and fast responses. |
| 🎭 VexPsyche (Emotions) | Simulated "free will". Vex has dynamically shifting moods (Stress, Boredom, Affection) that alter her personality and typing speed! |
| 🛡️ Rate Limit Armor | Graceful fallback and wait-timers when the API quota is exhausted, preventing app crashes. |
| 💾 Persistent Memory | All configurations, local intents, and chat histories are securely saved locally via IndexedDB (Dexie). |
| 🎨 Deep Customization | Change chat bubbles, switch avatars dynamically, and apply system-wide dark/light themes. |
| 🌐 Offline Mode | Uses a Naive Bayes classifier as a fallback to chat even without an internet connection! |
| 📦 Export / Import | Complete freedom to export and backup your database of custom synonyms and intents. |
Built with modern web and mobile frameworks to ensure blazing fast performance across all devices:
