Persistent cognition
at architecture level.
17 subsystems across 3 layers. Memory that compounds. Every decision verified. Every output passes through a constitutional ethics gate.
Free tier available. No credit card required.
17
Subsystems
3
Layers
100%
Auditable
<200ms
Latency
The Limitation
Stateless architectures cannot
produce persistent intelligence.
Current systems reset to zero after every interaction. They cannot accumulate knowledge, verify their own claims, or reason about time.
Memory
Resets every session
Persistent, compounding memory across weeks and months
Verification
No verification layer
Kernel-level truth verification during generation
Ethics
Alignment training only
Constitutional ethics gate on every output
Time
No concept of time
Temporal anchoring across all sessions
Transparency
Black box reasoning
Glass-box — every decision logged and auditable
Evolution
Static until retrained
Continuous evolution loops that compound every cycle
Architecture
The cognitive graph.
Every response passes through all three layers. Click any node to inspect its role in the cognitive chain.
Signal flow — every request traverses all layers before output
Benchmarks
Measured. Not marketed.
9 evaluation layers running continuously. Every score is auditable, every metric is live.
Cognitive Transparency
See inside the mind.
Every other AI is a black box. Primordia shows you exactly which subsystems fire, what evidence they find, and how decisions are made.
Ask Primordia anything. Watch which subsystems activate.
Capabilities
What persistent cognition enables.
Persistent Memory
Structured episodic and semantic memory that compounds indefinitely. The system builds genuine understanding of its operational domain across weeks and months.
Truth Verification
Kernel-level verification during generation. Every factual claim traced to provenance before output. Hallucinations caught at the architecture level.
Temporal Reasoning
Temporal anchoring across all sessions. The system understands when, in what order, and how events relate across time.
Constitutional Ethics
Configurable ethics gate on every output. Verifiable rules, domain-specific constraints, full audit trail.
Continuous Evolution
Each cycle compounds into the next. The architecture improves without retraining.
Glass-Box Transparency
Every reasoning chain, memory retrieval, and verification check is logged. Trace any output to its source.
Pricing
Choose your level of cognition.
Every tier includes persistent state, verified outputs, and constitutional ethics. The difference is depth.
Companion
Persistent memory. Real conversation.
- Echo conversation engine
- Mnemonic persistent memory
- Unlimited messages
- Memory across sessions
- API access
Storage
1 GB
API calls
10k/mo
Founding Operator
Full system access. Limited to 100 seats.
- All 17 cognitive subsystems
- Full governance & enforcement layers
- Memory, belief, and trace inspection
- Unlimited messages
- Priority access during system evolution
- Direct operator support channel
Storage
25 GB
API calls
250k/mo
Sovereign
Full architecture. No limits.
- All 17 cognitive subsystems
- Full consciousness architecture
- Dedicated support channel
- Advanced analytics & traces
- Unlimited messages
- Priority feature requests
Storage
25 GB
API calls
250k/mo
Enterprise
Dedicated infrastructure, custom integrations, SLA guarantee, on-premises deployment. Unlimited everything.
Compare Plans
See exactly what each tier includes.
| Feature | Companion | Founding Operator | Sovereign |
|---|---|---|---|
| Echo (Conversation Engine) | |||
| Mnemonic (Persistent Memory) | |||
| Kairos (Temporal Reasoning) | — | ||
| Aletheia (Truth Verification) | — | ||
| Simulon (Decision Simulation) | — | ||
| Astraea (Ethics Gate) | — | ||
| All 17 Subsystems | — | ||
| Reasoning Traces | — | ||
| API Access | Basic | Full | Full |
| Storage | 1 GB | 25 GB | 25 GB |
| API Calls | 10k/mo | 250k/mo | 250k/mo |
| Support | Community | Priority | Dedicated |
| Feature Requests | — | — |
Free Observer tier available — 20 messages/day with Echo. No credit card required. Try it free
Research
Published work.
Persistent Cognitive Architecture: A Three-Layer Framework for Stateful AI Systems
We introduce a novel architecture for AI systems that maintains persistent state across sessions through a three-layer cognitive framework. The core layer handles memory, verification, and temporal anchoring; the controller layer orchestrates cross-layer signals; the meta-executive layer provides ethics governance and self-reflection.
Aletheia: Kernel-Level Truth Verification for Language Model Outputs
We present Aletheia, a verification subsystem that operates at the kernel level of a cognitive architecture. Unlike post-hoc fact-checking, Aletheia traces every factual claim to its provenance during generation, achieving measurable reduction in ungrounded assertions without sacrificing output quality.
Astraea: Constitutional Ethics Gates for Autonomous AI Decision-Making
We propose a constitutional approach to AI ethics enforcement through Astraea, a subsystem that applies verifiable rule-based constraints to every output. Unlike alignment training, Astraea's constraints are explicit, configurable, and fully auditable — enabling domain-specific ethics enforcement at inference time.
Mnemonic: Compounding Memory Systems for Long-Horizon AI Reasoning
We describe Mnemonic, a memory subsystem designed for indefinite knowledge accumulation. Unlike retrieval-augmented generation, Mnemonic stores structured episodic and semantic memory that compounds over time, enabling AI systems to build genuine understanding of their operational domain.
FAQ
Common questions.
Get Started
The architecture exists.
Use it.
Persistent memory. Verified outputs. Constitutional ethics on every response. Start free, or claim one of 100 Founding Operator seats.
Free tier: 20 messages/day with Echo. No credit card.
Founding Operator — $99/mo
All 17 subsystems. 100 seats.