How we handle your data
A plain-language summary — written to be honest rather than exhaustive. A formal, legally reviewed policy is in progress; if anything here and that document ever disagree, we’ll fix it in the open.
Nothing is persisted. A guest journey lives in your browser tab and dies with it.
Your email (from the sign-in provider you choose), your journey state (so a conversation can continue across visits), your check-in throughline (the milestones and notes you choose to log), your subscription record if you have one, and a snapshot of each read we give you — the dimension estimates and the steps you mark done. These exist so the product works and remembers where you left off; they are service functionality, not research.
If you accept the “help tune the instrument” invitation, we’ll ask you — around 90 and 180 days after a read — how things actually went, and use your coded answer to make future reads more accurate. That’s the whole purpose. You can opt out at any time; declining is remembered and you’re never asked again.
No mood tracking, ever — there is no column for it. If the product detects a hard moment, that detection lives in memory for the session and is never written down, attached to your profile, or sent to analytics. Peer-space posts carry no account link of any kind — we couldn’t connect one to you if we were asked to. If you post with a peer “voice”, that voice is controlled by a key that lives only in your browser: we can’t connect a voice to your account either, which also means deleting your account cannot remove voice posts — that’s the honest cost of them being unlinkable.
Your conversation is processed server-side by an AI provider (Anthropic’s Claude) to produce the read and the advisor’s replies, gated on your consent. Prompts and scoring internals stay on our servers; the AI provider does not receive your email.
Everything keyed to your account can be exported or hard-deleted, completely — every table, both directions, verified by the same enumeration in code. Today these run as authenticated endpoints (/api/gdpr/export, /api/gdpr/delete); an account-menu surface for them is on the build list. Deleting removes research answers too, regardless of any consent you gave.
See also: How we handle safety. consuliq is in beta. Questions about any of this: the answer you get will be the true one.