The honest status page.
Most companies show you the roadmap and call it the product. This page shows the line between what is live today and what is still being built. It is linked from everywhere on purpose.
Where things actually stand.
Today Cintrel runs on a deterministic engine of seven rules plus one language-model synthesis pass per cycle, anchored by a persistent memory layer. The engine tick and the morning brief each run once a day on a fixed schedule: a tighter cadence is a hosting-tier change, not a code change. The eleven named agents below are how that capability is organized in the v1 spec, not eleven autonomous bots. They all run through that one shared engine; the one still in pack spec says so. The chips below are the honest current state.
No vapor, agent by agent.
The same patterns, wearing each industry's nouns.
A vertical is a noun map, a rulebook pack, and a pilot corpus. The engine, the rank function, the evidence rule and the refusal discipline ship unchanged. That is the point of them.
One pattern, many industries.
One engine underneath.
Every vertical runs the same codebase, the same schema, and the same caps. What changes is the rule pack and the noun map. A new industry is a configuration problem, not a rewrite, which is why the discipline in the engine matters more than any single rule.
The engine proposes. It does not execute. Every suggestion carries its evidence, confidence is capped in the schema, and a run with no resolvable evidence is refused rather than filled in. The detail on security, data handling, and continuity is on the trust page.
Pack specs above are designed, not built, and say so. If anything on this page ever reads as more certain than it should, that is a bug. Tell us and we will fix the wording.
Honest enough to bet on.
If radical honesty about what is and is not built is the kind of partner you want, let us show you the rest on your own stack.