Frequently asked questions

Can I use Mnemontron today?

Not yet. Mnemontron is pre-launch. There is no public API, no hosted service and no self-serve sign-up. Early access is by request.

How much does it cost?

There is no pricing yet. Mnemontron is not sold or offered as a paid service today.

What makes a Mnemontron answer different from a search result or a chatbot answer?

Every claim in an answer cites specific evidence: the content hash and version of the source, the span that supports it, when it was retrieved and its custody status. When the evidence is insufficient, Mnemontron declines to answer instead of guessing.

How do I know whether a person or an agent checked a claim?

Each claim records its verification separately. A claim verified by an agent is marked agent_verified; a claim verified by a person is marked as human-verified, with who verified it. The two are never merged.

Does it work with AI agents?

Yes, by design. Mnemontron includes a Model Context Protocol (MCP) server, an A2A agent card and llms.txt support, and signs its query and citation receipts so agents can verify them. These interfaces are not publicly available yet.

How is my data protected?

Sign-in uses passkeys only, with step-up re-authentication for sensitive actions. Evidence is encrypted at rest with per-tenant AES-256-GCM keys, tenants are isolated with row-level security, and an audit log is hash-chained. See the security page.

What happens when a source becomes outdated or disappears?

Mnemontron flags results whose sources are stale or missing, so you can see the problem. The flag does not silently change the ranking.

Can Mnemontron really store 100+ exabytes?

That figure comes from a capacity model, which is a projection. It has not been tested at that scale. The runs that have actually been measured are listed on the benchmarks page.

How do I get in touch?

A public contact address for mnemontron.com has not been published yet. It will be listed on the home page as soon as it is live.