Privacy Policy

Last updated 2 July 2026

Plain-English summary: We collect your email, the content you upload, and basic usage so the Service works and we can bill you. We don't sell your data. Your content is sent to Anthropic to generate answers — they don't train their models on it. You can export or delete everything anytime.

1. What we collect

Account data — your email address (for sign-in and billing) and, optionally, a display name.
Your Content — the sources, pages, and persona details you upload to build your twin.
Usage & cost metering — counts and token-based cost of the actions you take (chat, filing, organizing), so we can enforce plan limits and bill accurately.
Operational logs — minimal technical logs needed to run and secure the Service.

We do not run third-party advertising or cross-site tracking, and we don't collect more than we need to operate.

2. How we use it

To provide and secure the Service, generate your twin's answers, enforce plan limits, bill paid plans, respond to support requests, and meet legal obligations. We bill the actor who performs an action — usage is metered to the account that did it.

3. Who processes your data (sub-processors)

We rely on a small number of providers to run the Service. Your Content passes through these:

Provider	Purpose	Notes
Supabase	Database & authentication (hosting your account and content)	Data stored in the US
Anthropic	AI model that generates your twin's answers	Does not train its models on data sent through the API
Groq	Voice-to-text transcription	Audio is sent only when you use the microphone, solely to transcribe your speech
Vercel	Application hosting	Serves the app and API
Stripe	Payment processing (paid plans)	We never see or store your full card details
Brevo	Transactional email (confirmation & password-reset messages)	Receives only your email address, to deliver account emails
Cloudflare Turnstile	Bot protection on sign-up and sign-in	Runs a browser challenge; we never see the signals it evaluates

4. Who can see your twin

By default, visibility is controlled by you, per page. Owners see everything in their twin; guests you invite see only pages marked guest-visible; otherwise, people you haven't invited see nothing of your knowledge base — with one exception you control, described next. Tenant isolation is enforced at the database layer, not by the AI — the model never decides who may read what.

Community & public sharing. The Service includes an optional Community feed. Sharing there is always an explicit action you take, per note — nothing leaves your private knowledge base until you publish it. When you publish a note, you choose its audience: public (visible to any signed-in user in the Community feed, who may also ask your twin about that note), friends (visible only to accounts you've added as friends), or private (only you). You can also add other users as friends, which creates a mutual connection between your accounts and a shared activity surface. Publishing or adding friends does not change the visibility of anything you haven't explicitly shared, and you can manage what you've published.

5. Your rights — export & deletion

Export — download a full copy of any twin (every page as markdown) or your entire account at any time, from your twin's admin → Export / Delete.
Delete — delete a single twin, or your entire account (all twins, your guest access, and your profile), from the same page. Deletion is permanent and cascades immediately.
Access & correction — you can view and edit your content directly in the app. For other requests, contact us.

If you can't reach the admin page (for example, an account with no twin), email us and we'll action your deletion request.

6. Retention

We keep your content for as long as your account is active. When you delete a twin or your account, its content is removed from the live database promptly; routine backups, if any, age out on their normal cycle. Cost-metering records tied to billing may be retained as required for accounting and legal purposes, disassociated from deleted content.

7. Security

Each account's data is isolated by row-level security at the database, credentials never reach the AI or the browser, and components that act autonomously hold read-only access. No system is perfectly secure, but isolation is a first-class design goal rather than an afterthought.

8. Cookies

We use only the cookies needed to keep you signed in and operate the Service. No advertising or third-party tracking cookies.

9. Children

The Service isn't intended for anyone under 16 (or the age of digital consent where you live). We don't knowingly collect data from children.

10. Changes

We may update this policy; material changes are reflected in the "last updated" date and, where appropriate, announced in the Service.

11. Contact

Privacy questions or requests: vincent.jialun@gmail.com.

This policy is a good-faith, plain-language description of how the Service actually handles data — it is not legal advice. Before relying on it at scale, have it reviewed by a qualified privacy professional.