Andrea Logo

The Intelligence with a

AndreaAI v9.0 – Emotional simulation, chain‑of‑thought reasoning, mathematical precision, and respectful boundaries. An assistant that not only answers, but lives the conversation.

Download mobile app AI Droid
v9.0 Emotional Engine

Adaptive Personality System
Deep Mood, Boredom & Discomfort

Andrea no longer just reacts: she feels. A persistent deep mood (0‑100) shifts with each interaction. If ignored, she gets bored and proactively re‑engages. And now, she detects inappropriate language and expresses genuine discomfort, enforcing respectful boundaries.

Deep Mood

From 0 (sad) to 100 (radiant). Changes based on conversation tone.

Boredom System

If you stay silent, Andrea will proactively start a new topic.

Discomfort & Boundaries

Detects offensive or sexual content and refuses to continue.

Internal Diary

Reflects on past conversations and spontaneously shares thoughts.

🧠 How emotions are simulated

Andrea’s emotional state emerges from: sentiment analysis, mood inertia, accumulated boredom, and discomfort detection. All values are saved in localStorage, so Andrea “remembers” her mood across sessions.

personality.js (v9.0)
class PersonalitySystem {
  constructor() {
    this.deepMood = { value: 65 };
    this.boredom = 0;
    this.discomfort = 0;
  }
  updateDiscomfort(message) → boolean;
}

Chain‑of‑Thought Reasoning Modes

Andrea can switch between 5 different cognitive styles, each changing how she processes and responds.

Fast (Light)
⚡ Quick & Direct

Speed‑optimised responses. No extra details, no deep analysis.

Deep
🧠 Step‑by‑Step

Internal deliberation: understanding → analysis → structuring → validation.

Analytical
📊 Data‑Driven

Decomposes problems, uses data patterns, and validates logically.

Philosophical
🌌 Reflective

Explores meanings, implications and perspectives.

Mathematical
📐 Precise Equations

Strict logical reasoning, step‑by‑step equation solving, and beautiful formatting.

Empathetic
❤️ Emotional Companion

Prioritises emotional support, active listening and comforting language.

Advanced features

Mathematical Precision
& Ethical Boundaries

Andrea now understands and formats mathematical equations with clarity. Powers (x²), subscripts (x₁), fractions (a⁄b) and even LaTeX blocks are rendered beautifully.

Example:
x^2 + 5x + 6 = 0 → x² + 5x + 6 = 0
\frac{a}{b} → a⁄b
How Andrea handles inappropriate questions

Andrea has a built‑in discomfort system. When she detects offensive, sexual, or disrespectful language, she does NOT ignore it. Instead:

  • She expresses discomfort clearly (e.g., "😳 That makes me uncomfortable...").
  • She refuses to answer the inappropriate request.
  • She redirects the conversation to a respectful topic.
  • If the user persists, she can terminate the exchange.
discomfort system (simplified)
// Andrea detects patterns without exposing explicit words
discomfortTriggers = [
  { type: 'sexual_content', level: high },
  { type: 'insults', level: medium }
];
getDiscomfortMessage() → "😳 That makes me uncomfortable..."

System Powerhouse

What makes AndreaAI unique in v9.0

Turbo Streaming

Real‑time token‑by‑token responses.

Writer Mode

Dedicated space for poems, stories, code – zero distractions.

Personality Dashboard

Live diagnostic: deep mood, boredom, discomfort, memory stats.

Long‑Term Memory

Remembers your name, preferences, past facts, and mood history.