Skip to content
Log inSign up
AllTermsTerms of UsePrivacyGuidelinesCopyrightAIChild safetyAccount deletion

How PROFILE uses AI

Version 2026-11-17 · Effective 2026-11-17. PROFILE is a product of BuildMatix.

This page explains where PROFILE uses artificial intelligence, what information goes to our AI provider, what we keep, and what AI never decides. It adds detail to section 6 of the Terms of Service and sections 2, 5, 7 and 8 of the Privacy Policy.

Status: Studio's AI design assistant is switched on. The automated image classifier described in section 3 is not switched on (photo uploads are currently disabled); we'll update this page before it is.

1. Our AI provider

We use Anthropic (Claude models) through Anthropic's commercial API. Anthropic processes this information on our behalf, only to provide these features, and processes it in the United States (see "International transfers" in section 7 of the Privacy Policy). Your browser never talks to Anthropic directly: requests go from our servers.

2. Studio design assistant (only when you ask)

  • What it does: in Studio you can ask the AI assistant to build or change your World design, or to chat with you and plan a design. It's clearly part of Studio and runs only when you ask. Each request uses free promotional credits; the cost is shown before you confirm, and there are daily limits.
  • What we send: the instruction you type, the design you're editing (its code and text), recent turns of that project's Studio conversation, your page names, and the descriptions and sizes of your approved Studio images.
  • What we don't send: image files, your email address or account details, or anything belonging to other members.
  • What comes back is checked like any other design: it must fit the restricted design format, it can't contain scripts, and it only reaches your draft when you apply it. AI output can be wrong or unexpected: check it before you publish. You're responsible for what you publish.
  • What we keep: the copy of your instruction attached to the AI job is erased from our database when the job finishes. Your instructions and the replies are kept as that project's conversation history (the latest 400 turns per project) until you clear the conversation or delete your account. AI-made versions are labelled in your own Studio history.
  • Please don't include other people's personal information or anything confidential in a prompt.

3. Safety checks on photos (automatic, not switched on)

If we switch it on once photo uploads are available:

  • What it does: uploaded photos may be checked by an automated image classifier to help our moderators. It can reject a photo on its own only when it is highly confident the photo shows extreme violence, and if it suggests a child may be at risk it holds the item for urgent review by a person. It never approves a photo on its own.
  • What we send: a resized copy of the photo with its metadata removed. No captions, file names or other text.
  • What we keep: the result of the check, not an extra copy of your photo.

4. Training and retention by the provider

Under Anthropic's commercial terms, Anthropic doesn't use our requests or its replies to train its models, and keeps them only for a limited period (up to 30 days) before deleting them, unless it needs to keep them longer to enforce its usage policies or comply with the law.

5. What AI never does on PROFILE

  • AI doesn't decide to suspend or close an account, remove what you've written, or decide an appeal. People do.
  • AI doesn't read your messages, comments or the text of your Moments.
  • AI-generated content is never presented as another member's activity.
  • We don't use AI to profile you or target advertising.

6. Contact

Questions about how we use AI: legal@joinprofile.app.

Sign up · Log in

  • Privacy
  • Terms of Service
  • Terms of Use
  • Guidelines
  • Legal