Privacy Policy

SF2X is a small, independent project — built and run by one person. This policy explains, honestly, what we collect, why, who it's shared with, and the real, self-serve controls you have over it. Two things shape everything below: we never sell your data, and we never train any model on your private conversations.

1 · Talking to her without an account

You can use the public brain at sf2x.com without signing up. As a guest:

• No account, no profile. We don't create an account or build a profile about you. You talk to her shared brain — you can't read anyone's private memory, and nothing you say is written into her long-term memory.
• Your message is processed by OpenAI. To generate a reply and run an automated safety screen on it, each message is sent to OpenAI as a processor (under OpenAI's API terms, which do not use API content to train OpenAI's models). Please don't share sensitive personal information you wouldn't want processed this way.
• Your IP is used transiently for safety, not identity. We use your IP address only to rate-limit abuse and protect the service — never to identify, track, or profile you. It is not stored as part of any profile.
• Your transcript stays in your browser. A guest conversation lives in your own browser's local storage, on your device. Clear your browser data to erase it — we don't keep guest transcripts on our servers.

2 · What we collect if you create an account

Information you give us

• Email address — to sign in. SF2X is passwordless: you receive a one-time 6-digit code by email. We never ask for, store, or see a password.
• Optional birth data for the Origin Signature — date, and optionally time and a coarse location. You choose whether to provide this; it is used only to compute your symbolic Origin reading.
• Consent records — timestamps of the choices you make (agreeing to the Terms, opting in to memory). Memory is only stored if you opt in, and that opt-in is enforced in our code, not just promised here.
• What you say to her, and content you create — stored privately in your own memory lane (see §4) when you've consented to memory.

Information collected automatically

• A session cookie (sf2x_session) — an httpOnly, Secure cookie that keeps you signed in. It is strictly necessary, not an advertising or tracking cookie.
• Operational logs — minimal request/security logs (e.g., IP for rate-limiting, error traces) used to run and protect the service. We do not run third-party ad or analytics trackers that profile you.

What we do not collect

Passwords (there are none) · advertising identifiers · third-party tracking profiles · full payment-card numbers (if you ever buy something, card details go straight to Stripe and we never see them).

3 · How we use your data — and the legal basis

We do not use your data for advertising, for sale to third parties, or to train our own models. You can withdraw consent at any time (see §7); that doesn't affect processing already done.

4 · Your memory & how the brain learns

This is the part most products get wrong, so we'll be exact:

• Your memories live in a private lane that is yours alone. Each member's cells are stored under a cryptographically-scoped owner key (mem_<your id>). Another member cannot read, search, or reach your lane — this isolation is enforced server-side and verified by an automated cross-tenant test, not just asserted here.
• The shared "world brain" is not built from your private words. Her public knowledge is grown from openly-licensed sources and material the operator curates — your private conversations are never mined into the shared brain, and are never used to train any model.
• Memory is opt-in. If you don't consent to memory, she talks with you without writing durable cells about you.

5 · AI processing & outputs

• Processors for AI. Your messages (and, for members who opted in, the text we store) are sent to OpenAI to generate replies, create the numeric "fingerprints" (embeddings) that power recall, and run safety moderation. OpenAI processes this as our sub-processor and does not use API content to train its models.
• Outputs can be wrong. AI replies may be inaccurate. Anything symbolic — the Origin Signature, planetary or "field" readings — is a low-confidence lens for reflection, not a fact and not professional advice. Use your own judgment.
• No solely-automated decisions with legal effect. SF2X does not make automated decisions that produce legal or similarly significant effects about you.
• Human review is minimal and safety-only. Because this is a one-person operation, the operator may review content that is flagged for safety or abuse, to keep the service and its users safe. We don't read your private memory for any other purpose.

6 · Who we share data with

Only the providers that run the service, each strictly to perform its function. We do not sell your data and do not share it for advertising.

We may also disclose data if required by law, or to protect rights, safety, and the integrity of the service.

7 · Your rights — and the real buttons that honor them

Most policies promise rights and bury them behind an email. SF2X ships them as one-click tools:

• Export everything (data portability). A self-serve export returns your profile, your memory cells, and their connections in a machine-readable form — your data, in your hands.
• Delete everything (erasure). A self-serve "forget me" performs a hard deletion of your account, memories, connections, and Origin data — not a soft hide. We then write a tamper-evident, PII-free ledger entry that proves the erasure happened without keeping any record of who you were.
• Access, correct, withdraw consent, pause learning, or delete any single memory — directly, or by emailing us.

GDPR (EU/UK): access, rectification, erasure, restriction, portability, objection, and the right to lodge a complaint with your supervisory authority. California (CCPA/CPRA): know, delete, correct, and opt out of "sale"/"sharing" — which we do not do — with no discrimination for exercising your rights. To make any request, email campiper84@gmail.com. We aim to respond within the timeframe the law requires.

8 · Where your data is processed

SF2X is operated from, and most processing happens in, the United States (OpenAI, Supabase, Cloudflare's global edge). If you are in the EU/UK and your data is transferred to the US, it is done under appropriate safeguards such as Standard Contractual Clauses, as offered by our processors. (Specific transfer mechanisms will be confirmed with counsel and updated here.)

9 · Retention, deletion & security

Retention. Guest conversations are never stored on our servers. Member data is kept while your account is active and is deleted promptly when you erase it (we may retain limited records we are legally required to keep, e.g., for tax or to evidence a deletion). Operational logs are short-lived. Backups are rotated and purged on a rolling basis.

Security. Passwordless sign-in (nothing to leak) · cryptographically-scoped per-member isolation, verified by an automated test · row-level security and service-role-only data access · TLS in transit · least-privilege keys held only server-side · hard rate-limits and a daily spend ceiling to blunt abuse. No system is perfectly secure; if a breach affecting you occurs, we will notify you as the law requires.

10 · Children

SF2X is not directed to children under 13 (or the higher minimum age in your region, e.g., 16 in parts of the EU) and we do not knowingly collect their data. If you believe a child has provided us data, contact us and we will delete it.

11 · Changes & contact

We may update this policy; we'll change the version and effective date above and, for material changes, give notice (e.g., on the site or by email). Questions or requests: campiper84@gmail.com.