Skip to content

Policy updates

When we update our privacy policy, terms of service or refund policy, we list it here with the date and a summary. The newest is at the top.

  1. Privacy Policy 1.09, Terms of Service 1.6, Refund Policy 1.1

    We've updated our privacy policy, our terms of service and our refund policy. The new versions took effect on 9 November 2026.

    What changed in Privacy Policy 1.09

    • We don't train AI models on your information at all. We've removed the line that said information used to improve an AI feature might stay in a trained model.
    • We've switched Help improve indi off for some accounts that hadn't chosen either way, so it's worth checking it's set the way you want. It decides whether your child's record, your conversations, your notes, your documents and your account details are included in the analysis we use to judge how indi is performing. Turning it off leaves your information out of the next set we draw together, though information already drawn together may still be analysed.
    • A new "Automated decisions" section. We don't use AI or other automated systems to make decisions that significantly affect you or your child. A person at indi makes the call on our automatic checks of chat replies and Letters of Support, and you can ask a person to look again at anything those checks did.
    • We corrected what we said about where a chat reply is reviewed. The decision and its record are still made inside indi, but a reviewer may also look at the technical record of the conversation held by our monitoring services.
    • We set out the legal reasons we rely on to use your information, and listed the privacy regulators you can contact if you're unhappy with how we handle your information.
    • Ending a subscription stops paid access and deletes nothing. Deleting your account is your choice, at any time.
    • Crash reporting stays on whatever your analytics setting, so we can keep indi working and secure.
    • A new section on marketing emails: what we send, that we never use your child's information to decide it and how to opt out.

    What Terms of Service 1.6 and Refund Policy 1.1 say

    • Nothing you or your family put into indi is used to train, fine-tune or build an AI model. There are no exceptions.
    • If we make a material change to the Terms that adversely affects your rights or obligations, we'll give you at least 30 days' notice before it takes effect.
    • Cancelling doesn't delete your data. You can download it for at least 30 days afterwards, and it's still there if you come back.
    • For clinics: how Clinic Plans, seats, credits and invoices work, that a signed agreement with us applies alongside the Terms and that indi is not a medical device.
  2. Privacy Policy 1.08

    • When a clinic asks us for a demo or a report on our website, our server also tells Meta about the request. It sends the visitor's IP address, browser details and Meta's ad click cookies. It only does this if they accepted marketing cookies.
  3. Privacy Policy 1.07

    • The advertising cookies on our website also record that a clinic asked us for a demo or a report through one of our forms. They never record what the clinic wrote.
  4. Privacy Policy 1.06

    • We named Cloudflare as the provider that delivers and protects our website and web app, and described the page analytics it collects without cookies.
    • We added session replay to the usage data we collect: when recordings are made, and that text, images and anything you type are hidden on your device before a recording is sent.
    • We added Cloudflare and its page analytics to what we say about sending information overseas.
  5. Privacy Policy 1.05

    • We described how we share information with analytics, monitoring and observability providers, and with third parties generally.
    • We made clear that safety and quality checks on our AI features use your information as you entered it, not a de-identified copy, and that they carry on whether or not you take part in model improvement.
    • We set out which AI responses a person reviews: ones you tell us were unhelpful or wrong, and ones our automated safety checks flag. The reviewer may read the surrounding conversation to see how a response was reached.
    • We said a review creates a record that's deleted with the conversation, and that the review happens inside indi.
    • We named the countries your information is sent to, the basis for sending it and what we don't send overseas.
    • We updated where our mobile attribution provider processes information, from its published list of subprocessors.
    • We pointed to the Cookie Policy for the advertising cookies on our website.

Earlier versions of our policies aren't published, so these entries describe them rather than link to them.