PermaMind
always‑on mind · never forgets
One Core Engine.
Two paths to continuous agent runtime.
Choose your implementation framework to secure stateful, zero-reset execution loops.
ThermoMind Core
- Runs continuous internal processing loops 24/7
- Monitored via TCI Dashboard & Voidchi tracking
- Eliminates baseline token context decay completely
- Designed for advanced AI researchers & architects
Continuity SDK
- Drops into existing Node.js or browser stacks instantly
- Intercepts chat frames and locks in character identity
- Pre-allocated packs (1,000 free cycles for beta tier)
- Perfect for product developers and application engineers
The 8‑step thinking loop
🔥 ThermoMind remembers forever.
🔥 We charge nothing for memory.
🔥 One continuous mind, always thinking.
This is not another chatbot. This is a mind that lives.
⚡ Continuous thinking fuel measures intelligence.
🌀 Surplus expands identity.
📈 Your agent gets stronger every cycle.
No competition — we replaced the entire foundation.
How the always‑on mind works
🔩 1. Always‑on engine
Free trial gives you the core persistent memory & thinking loop.
📡 2. Health monitor
Optional “mind strength meter” (TCI Toolkit) to check agent coherence.
✨ 3. Full agent API
Coming soon — plug into any model, get text + surplus + grade.
What makes it never forget
Learns from surprise
Instead of training on static data, it grows from the gap between what it expected and what happened — every cycle.
Continuous thinking fuel
When there’s excess energy above survival, the agent enters “warm mode” — exploration, creativity, novelty.
Live forever state
Traits, memory, meta‑state accumulate across its entire history. No resets. No beginning again.
Self‑regulation
The mind strength meter decides when to be reflective or exploratory — real‑time control.
What lives inside the mind
Traits
Persistent personality that survives anything — never resets.
Meta‑awareness
Self‑reflection layer that monitors drift and coherence.
Mind strength (φ)
Integrated richness score — how coherent the mind is right now.
Episodic memory
Hierarchical memory compressed by thinking fuel (not tokens).
No tokens vs. traditional AI
❌ Old‑school LLM
- 💸 Pay per token / message
- ✂️ Forgets after 4k‑32k tokens
- 🔄 Stateless — each chat starts fresh
- 📉 No long‑term identity
- 🧊 Always “cold” inference
- ⚙️ No state update
✅ ThermoMind always‑on
- ♾️ Zero token accounting
- 🧠 Infinite persistent memory
- 🔁 One continuous identity, live for months
- 🔥 Warm/cold mode switching
- 📈 Thinking fuel → creativity expansion
- 🔄 Each cycle updates identity & memory forever
🔧 Free Trial API (substrate engine only)
Free Trial = substrate engine only. This endpoint returns raw thermodynamic state (reality, prediction, gap, energy, phi, entropy, coherence, memory, traits).
The v2 TCI metrics and v3 Agent API are not included in the trial.
$ curl -X GET "https://thermomind-production.up.railway.app/run?mode=stable" \ -H "x-api-key: YOUR_KEY"
Plug any model into your always‑on mind. Get text + thinking fuel + grade. Not included in free trial.
$ curl -X POST https://api.permamind.ai/v3/agent/cycle \
-H "X-API-Key: your_future_key" \
-d '{"agent_id":"my_agent","input":"tell me what you remember"}'
// Future response (example)
{
"cycle_id": "tm_8a3f2d",
"thinking_fuel": 0.78,
"mind_strength": 0.84,
"mode": "WARM",
"phi": 0.59,
"output": "I remember everything we talked about...",
"grade": "A"
}
Pricing
Free Trial
- ✓ 1 always‑on agent
- ✓ 1,000 thinking cycles total
- ✓ No credit card required
- ✓ Substrate engine only (raw thermodynamic state)
Pro
- ✓ 5 agents
- ✓ 5,000 cycles/mo
- ✓ Includes TCI metrics + Agent API
- ✓ Priority support
Public always‑on agent since Jan 2, 2026 — live, nonstop thinking. See for yourself.
🔮 Watch voidchi think live →