What we can see, and what we can't.

The second brain is ordinary text files in a private repository in your name. We don't host it.

If you stop working with us, everything stays with you and keeps working exactly as before. No lock-in, no export, nothing to ask us for.

The system doesn't send us any information automatically. After the first month of support we remove our access, unless you ask us to keep it.

If you choose monthly support and a particular check needs access, we ask you for it each time.

During setup and the first month of support we have access to your private repository, to what you approved in the mapping call.

At the end of the first month we remove our access by hand, unless you ask otherwise. We tell you when it's done.

You can remove us from the repository at any moment, yourself and without asking. We don't need to approve it.

The files of the system we build for you sit in a private repository in your name, and Claude reads them through your own subscription with Anthropic. We don't pass their content to anyone.

We don't read content without an operational reason and without you asking. We don't use your content to train models. What we learn about how the system works, without your content, becomes improvements that reach all clients.

We don't name you as a client and don't quote you without written permission.

The questionnaire on this site passes information to a few services, and this is how it works:

Anthropic (Claude): to write the short summary at the end, your first name and a few of your answers are sent to Anthropic through our account.

Saving your answers: if you leave an email, your name, answers and email are saved with us so Alon can get back to you. Without an email, your answers aren't saved with us, and the draft stays only in your browser.

Notification: when you leave an email, Alon gets a message with your name and email.

Hosting: the site is hosted with a hosting provider, and requests to the site are recorded in its system logs.

To delete your answers, ask at alon@ak-studio.co.il.

The system works with Claude through your own subscription, in your name. Your relationship with Anthropic is directly with them, under their terms, and we are not a party to it.

In practice: what you send to Claude is processed by Anthropic under their terms. The access you give us during setup is a separate permission, and you can remove it at any moment.

The setup includes what's listed on the pricing page, including the first month of support. What isn't listed there isn't included, and that's deliberate.

Requests beyond that are separate work, and we price them before we start. We won't do extra work and send a bill afterwards.

The Claude subscription, about $20 a month, is paid by you directly to Anthropic.

The setup is $390, paid up front, and includes the first month of support. If after the mapping call you or we decide it isn't a fit, we refund the full payment and don't start building.

At the end of the first month we ask whether you want to continue with monthly support, $49 a month. It starts only if you approve, with no commitment period. You can stop it with a message, and the month you've already paid for runs to its end.

Invoices are issued properly through our system.

Keeping your own accounts safe: passwords, access, and a backup somewhere convenient. We'll show you how, but it stays with you.

What goes into the system is your responsibility. If you put your own customers' information into it, your obligations to them still apply.

We build the system carefully and we commit to fixing faults that come from our work. We aren't liable for business decisions you made on the basis of AI output.

AI models get things wrong. The system exists to give them better context, not to make them an authority.