Consumer Product
⟡ Active Mirror
Reflective AI that asks questions to help you think — not chatbots that tell you what to do.
Try it live →
Core Features
⟡ AI Twins
Four cognitive companions — Guardian, Scout, Synthesizer, Mirror — each with distinct thinking styles.
◈ BrainScan
8-question cognitive architecture quiz. Discover your archetype and get matched with your ideal AI Twin.
△ Mirror Proof Protocol
Cryptographic attestation of human attention. Proof you actually read the consent.
◇ Sovereign Mode
On-device AI via WebLLM. Your data never leaves your device. True privacy.
Cloud vs Sovereign
| Feature | Cloud Mode | Sovereign Mode |
|---|---|---|
| Model | Llama 3.3 70B (via Groq) | Phi-3.5 Mini (WebLLM) |
| Speed | Fast (~1s) | Slower (~5-10s first load) |
| Privacy | Data sent to cloud | 100% on-device |
| Offline | Requires internet | Works offline (after download) |
| Model Size | N/A | ~2GB download |
The Four Twins
Guardian
Protects your focus. Filters distractions, maintains boundaries, keeps you aligned with your goals.
Best for: Architects, Builders, Strategists
Scout
Explores possibilities. Finds connections, surfaces opportunities, ventures into unknown territory.
Best for: Explorers, Connectors
Synthesizer
Connects the dots. Weaves ideas into frameworks, finds patterns, creates coherence from complexity.
Best for: Analysts, Creatives
Mirror
Reveals blind spots. Asks uncomfortable questions, illuminates shadows in your thinking.
Best for: Scholars
Cognitive Archetypes
BrainScan identifies one of 8 cognitive archetypes:
| Archetype | Tagline | Matched Twin |
|---|---|---|
| Architect | Systems thinker. Pattern builder. | Guardian |
| Explorer | Boundary pusher. Unknown seeker. | Scout |
| Builder | Maker of things. Shipping machine. | Guardian |
| Analyst | Data whisperer. Truth finder. | Synthesizer |
| Connector | Bridge builder. Network weaver. | Scout |
| Creative | Idea generator. Pattern breaker. | Synthesizer |
| Scholar | Deep diver. Knowledge keeper. | Mirror |
| Strategist | Move maker. Future seer. | Guardian |
Consent & Safety
⚠️ Active Mirror is experimental AI. It is not therapy, medical advice, or a substitute for professional help. Results are for entertainment and reflection only.
Mirror Proof Protocol
Before using AI features, users must complete informed consent:
- Full Mirror Proof — 3 acknowledgments + cryptographic hash (for AI chat)
- Light Consent — Quick banner acknowledgment (for quizzes)
Consent is stored per-session and carries across all pages once granted.
MirrorGate Protection
All AI responses pass through MirrorGate, a safety layer that blocks harmful content before it reaches users.
⟡ Try Active Mirror: activemirror.ai