What Perren learns about your business is stored on your computer, not on ours. The facts she remembers and the full record of your conversations live in a file on your own machine. We cannot read them, and they are never uploaded to us.
To answer a question she has to send that question — and the records needed to answer it — to a language model over the internet. That is the one thing that leaves your computer, and this page explains exactly what it is, where it goes, and what is kept.
Perren is made by Tyrus Labs LLC, a New Jersey limited liability company. In this
policy, “we” and “us” mean Tyrus Labs LLC, and “Perren” means
the Perren desktop application and the licence service at api.perren.ai that supports it.
Perren is a desktop application. The following are written to a database file on your own machine and are never transmitted to us:
Erasing Perren's memory from within the application is permanent. It clears both what she knows and the conversation log, and we are offered no copy of either, because we never had one.
Perren does not run a language model on your machine. When you ask her something, the application
sends, over an encrypted connection, to our service at api.perren.ai, which forwards it
to Anthropic:
This is not optional and it is not incidental — it is how the answer is produced. If your business information is confidential to a degree that it must not be processed by a third-party model provider at all, Perren is not the right tool, and we would rather say so here than bury it.
We do not store any of it. Our service passes the request through in memory and keeps no copy of the question, the records sent with it, or the answer returned. Anthropic processes the request under our commercial agreement with them; under that agreement your inputs and outputs are not used to train their models. Their handling is described at anthropic.com/legal/privacy.
When Perren searches the web — either because you asked her to, or because she offered and you accepted — the search query is sent through our service to Brave Search. Your records are not sent with it. Brave's handling is described at brave.com/privacy.
Each request carries a signed licence token identifying the subscription it belongs to. It is verified by arithmetic on our side and contains no personal information beyond that subscription identifier.
Our service keeps three things, and this is the complete list:
| What | Why | How long |
|---|---|---|
| A daily total of tokens used, per subscribing organization | To operate fair-use limits and bill correctly | Deleted automatically after 30 days |
| The time each licence seat was last used | To count how many seats are active | Retained while the subscription is active |
| Which services you have connected, and a count of calls to each | To know which integrations to support, and to bill usage correctly | Retained while the subscription is active |
All three are counts. None of them contains your questions, Perren's answers, your business records, your name, or your email address. The third records only that a connected service was called and how often — never what was asked of it, and never what came back. We do not maintain server-side logs of prompt or response content, and we do not operate analytics or advertising trackers inside the application.
Payments are processed by Stripe. Stripe collects the billing details needed to take payment; we never receive or store your full card number.
If you choose to pay by bank debit (ACH) instead of card, Stripe's Financial Connections lets you link your bank account by signing in to your bank. In that flow Stripe gives us a token standing in for your account and routing number, and the name and address on the account. We use those only to set up the payment method and to confirm the account belongs to you. We do not receive your bank login, and we do not request or receive your balances or your transaction history. You can disconnect the account at any time.
We keep the business name, billing contact email and subscription status needed to run the account, for as long as the account exists and for as long afterwards as tax and accounting law requires.
We keep the correspondence so we can help you, and no longer than we need it.
Perren can be connected to accounts you already use — for example Google Workspace or Microsoft 365 — so that she can read the documents you point her at and create the ones you ask for. Connecting an account is always your explicit choice, made through that provider's own consent screen, and it can be undone at any time.
Who carries out the connection. We use Composio, a specialist integration provider, to perform the connection and to make the calls you ask Perren to make on your behalf. Your request and the data it returns pass through their service. They act only on our instructions, we do not store that content on our servers, and none of it becomes part of anything we keep — but they are in the path, and you should know that before connecting an account that holds your finances. Depending on the provider, the access token for a connected account is held by Composio or supplied by us at the moment of each call.
This is the one place where something other than a question to a language model leaves your computer, so we would rather name it plainly here than have you discover it. It applies only to accounts you have explicitly connected. Nothing in Perren's memory, and nothing in your conversation log, is ever sent to Composio.
Google user data. Perren's use and transfer of information received from Google APIs to any other app will adhere to the Google API Services User Data Policy, including the Limited Use requirements.
We do not sell your information and we do not share it for advertising. The only third parties involved are the ones needed to make the product work, each named above:
We may also disclose information if the law requires it, or to protect our rights or someone's safety. If our business or its assets are ever sold, account records may transfer with them; what is on your computer is yours and does not.
Depending on where you live — including under the California Consumer Privacy Act and the UK and EU General Data Protection Regulation — you may have the right to access, correct, export or delete personal information we hold about you, and to object to or restrict certain processing. You may exercise any of these by writing to the address below, and we will not treat you differently for asking.
In practice, most of what you might want is already in your hands: the substance of what Perren knows about you is on your computer, and you can export or erase it yourself from inside the application, without asking us. What we hold is the short list in section 3.
We have not sold or shared personal information for cross-context behavioural advertising, and we do not intend to.
Traffic between Perren and our service is encrypted in transit. The licence signing key and the model provider key are held by us and never sent to the application — which is why the application talks to our service rather than holding a key of its own. Access to our systems is limited to people who need it. No system is perfectly secure, and we do not claim otherwise.
We are based in the United States and our service providers process data in the United States and other countries. Where personal information is transferred out of the UK or EEA, we rely on the appropriate safeguards offered by those providers, including standard contractual clauses.
Perren is a business tool sold to organizations. It is not directed at children and we do not knowingly collect personal information from anyone under 16.
If we change this policy we will update the date at the top of the page. If a change materially affects how your information is handled, we will tell subscribers directly rather than relying on you to re-read the page.
Tyrus Labs LLC, a limited liability company organized in New Jersey, United States.
Privacy questions and requests:
privacy@perren.ai
Our postal address is available on
request, and we will provide it to any regulator or data subject who asks.