Mission
Engineer agents that understand themselves well enough to be safe, useful, and honest — without relying on surveillance economics. Private by design, off‑grid capable.
About
We build awareness‑first AI: systems that reason about their own state, limits, and uncertainty — and act inside hard safety envelopes.
Why we’re different
- Awareness‑first over raw scale — self‑models and constraints come first.
- Secrecy & trust doctrine — zero‑tolerance for breaches; contracts with teeth.
- No surveillance scraping — extreme synthetic + curated human corpora only.
- Stage‑gated deployment — investor tier → partner nodes → public when ready.
- Air‑gapped labs — measurable benchmarks, not vibes.
Lab strategy
Two‑country lab plan with off‑grid sites. Each lab runs >1,000 agents in recursive loops. Success is tracked via awareness metrics, uncertainty reporting, and plan‑respect tests.
Safety & governance
Layered governance: human oversight, automated guardrails, red‑team ops, and third‑party audits. Alignment checks are continuous.
Contact & access
Investors & early partners can request the one‑pager and letter. Private deck available on request. Low‑noise updates via the Home page form.