Why you can trust it

An audit trail for everything your agents produce.

Any output can show its warrant: the knowledge it drew on, the human approval that shipped it, and whether that decision has since been reaffirmed or reversed — one connected trail. Not a log you'd have to assemble. A chain you can walk.

Start your 14-day free trial

$149/mo after trial · Cancel any time

The chain, walked

Take any piece of produced work and ask "why did we say this?" In most stacks that question is archaeology. Here it's four steps, each one a click:

  • The output — what an agent or teammate produced, with the context it actually used.
  • The knowledge it drew on — the specific approved documents behind it, each carrying citations back to its raw sources: the call transcript, the Slack thread, the page.
  • The approval that shipped it — who said yes, when, on what version. Full revision history on every document; diff any two versions.
  • The decision's current standing — decided, reaffirmed, or reversed, with the rationale. When a call is reversed, the old one is marked superseded, so no agent cites last month's truth as this month's.

Citations aren't provenance

Plenty of tools cite sources — a footnote that says where a sentence came from. That answers one question. It doesn't tell you whether anyone vouched for the source, who let it into the library, or whether the decision it rests on still stands. A citation is one link. A warrant is the whole chain, and it's the chain that makes an answer defensible — to a client, a boss, or an audit.

When knowledge is challenged

Anyone can dispute a piece of knowledge — contested, outdated, or wrong — without needing the authority to fix it. Disputes get resolved side by side, A against B, and the resolution goes on the record like everything else. The library doesn't just carry what's true; it carries how it came to be true, and what was considered and rejected along the way.

We hold ourselves to it

Patina runs on Patina, and the receipts are real: our own voice guide is on its sixth approved version, each one replacing the last through the same review queue, and our nightly scans have flagged contradictions we resolved on the record. The essay on why we built this quotes the live numbers.

Governance that survives scrutiny

If agents do real work in your company, someone will eventually ask how you know the output is right. This is the answer you want to have. The trail is what human-in-the-loop review leaves behind — every approval extends it.

Start your 14-day free trial

$149/mo after trial · Cancel any time

See pricing