Does Clue train on your data?
Clue built its own foundation model for its AI features and says so. What that model was trained on, the policy does not say — though it is clear no data reaches Anthropic.
This clause is real, but it doesn't resolve the training question - see the finding below.
"Through AWS Bedrock Clue operates its own Foundation Model of the LLM, which is a basic version of the LLM trained by Clue for Clue's AI features."
- 1 Oct 2026
- 5 Oct 2026
- All accounts (free and Clue Plus)
- global
- Fetched from the company
- D7063AF0…54A2
Why this is still UNCLEAR: The AI features are opt-in on explicit consent under Article 6(1)(a), and the policy is unusually careful about where the data does not go: "No data is shared with the operators of other Foundation models or the provider of the LLM", and "No data is transferred to Anthropic". That is better than most of this category manages. The gap is in the quoted sentence. Clue operates its own Foundation Model "trained by Clue" and never says what it was trained on. For an app holding cycle dates, symptoms and sexual activity, the difference between processing that data and learning from it is the question, and consent to the first is not consent to the second. Two passages come close enough to be worth naming. Records of AI interactions — "user queries, contextual information used to generate responses, evidence packets and generated outputs" — are retained and used to "operate, maintain, secure and improve AI functionality, identify failure modes, evaluate answer quality and develop future AI-supported functionality". And where information derived from those interactions "is used to improve predictions, develop new algorithms or create more valuable features", the policy routes it into its health analytics purpose, whose consent text reads: "I agree to the health data I track in the app being processed for analytics purposes so Clue can improve my predictions, develop new algorithms and create more valuable features with personalized insights." That processing is de-identified, kept on EU servers, and happens only "if you have toggled this on in your privacy settings". Developing algorithms from health data you opted into is not nothing — but it is not a statement that your data trains the Foundation Model, which is the question this entry is unclear about, and Clue's own careful drafting elsewhere shows it can say so plainly when it means to.
What each plan can refuse
The same questions, asked of every plan we could identify. Every cell comes from a quoted clause or a documented search - see the evidence above.
- Not established
- Not stated
- —
- —
- Not established
How to opt out of Clue AI training
Plan by plan, from what Clue's own policy says about refusing. Where it publishes the steps, they're listed here with the limits they carry. Where there's nothing to switch off, or nothing graded, this says so instead of inventing a menu path.
All accounts (free and Clue Plus)
Not graded, so no route is pointed at. The AI features are opt-in on explicit consent under Article 6(1)(a), and the policy is unusually careful about where the data does not go: "No data is shared with the operators of other Foundation models or the provider of the LLM", and "No data is transferred to Anthropic". That is better than most of this category manages. The gap is in the quoted sentence. Clue operates its own Foundation Model "trained by Clue" and never says what it was trained on. For an app holding cycle dates, symptoms and sexual activity, the difference between processing that data and learning from it is the question, and consent to the first is not consent to the second. Two passages come close enough to be worth naming. Records of AI interactions — "user queries, contextual information used to generate responses, evidence packets and generated outputs" — are retained and used to "operate, maintain, secure and improve AI functionality, identify failure modes, evaluate answer quality and develop future AI-supported functionality". And where information derived from those interactions "is used to improve predictions, develop new algorithms or create more valuable features", the policy routes it into its health analytics purpose, whose consent text reads: "I agree to the health data I track in the app being processed for analytics purposes so Clue can improve my predictions, develop new algorithms and create more valuable features with personalized insights." That processing is de-identified, kept on EU servers, and happens only "if you have toggled this on in your privacy settings". Developing algorithms from health data you opted into is not nothing — but it is not a statement that your data trains the Foundation Model, which is the question this entry is unclear about, and Clue's own careful drafting elsewhere shows it can say so plainly when it means to.
What this policy has done since we started watching
1 change detected since 30 Aug 2026. Each one is a diff between two archived copies, and each links to the passage that moved.
Everything we read
Vendor response
We haven't yet written to Clue about this verdict, and they haven't disputed, clarified or announced a change to any clause quoted above. When either happens it's published here, word for word.
Report a wrong clause