legal

AI Policy

Last updated: March 2026

Noodle uses artificial intelligence to power Untangle mode — a feature that surfaces connections between your captured thoughts. This policy explains how that AI works, what data it processes, and the principles we follow.

We believe in AI that is transparent, optional, and in your service — not the other way around.

1. What AI does in Noodle

Untangle mode

Untangle is an optional AI feature available on Pro and Teams plans. When you activate Untangle, our AI analyses a selection of your recent thought entries to identify patterns, themes, and connections you may not have consciously noticed. The output is a map of linked ideas and a plain-language insight summarising what it found.

Untangle is always triggered by you — it runs when you explicitly open the Untangle view. It is never running passively in the background without your knowledge.

Gentle nudges

Noodle may occasionally surface an older thought entry that appears related to something you recently captured. This connection is identified by our AI using the same process as Untangle. You can disable nudges at any time in Settings.

What AI does not do

It does not read or process your thoughts unless you explicitly trigger an AI feature.
It does not send your data to advertisers or third parties for commercial purposes.
It does not make decisions about you — all AI outputs are suggestions for you to consider and accept or ignore.
It does not use your content to train AI models. See Section 4 below.

2. What data is processed by AI

When you use Untangle or receive a nudge, a limited selection of your thought entries (typically the most recent 20–50 entries, or those you have highlighted) is sent to our AI provider for processing.

This data is:

Transmitted over encrypted connections (TLS).
Processed in memory only — our AI provider does not store your content after processing.
Associated with a pseudonymous session identifier, not your name or email.

Voice notes are transcribed locally on your device before any AI processing. Raw audio is never sent to our AI provider.

3. Our AI providers

We work with third-party AI model providers to power Untangle. Our current provider is Anthropic (Claude). By using AI features in Noodle, you acknowledge that relevant content will be shared with Anthropic for the sole purpose of processing your request.

Anthropic's data handling is governed by their usage policies, which prohibit training on API data by default. We will update this section if we change our AI providers or their policies change materially.

4. We do not train AI on your data

Noodle does not use your thought content to train, fine-tune, or evaluate AI models — whether our own or those of our providers. Your thoughts belong to you. They inform what AI tells you about yourself; they do not become part of any model's knowledge.

We contractually require our AI providers to uphold the same restriction.

5. AI is always optional

All AI features in Noodle are opt-in. You can:

Use the entire app without ever activating Untangle.
Disable AI nudges in Settings → Notifications.
Disable all AI processing entirely in Settings → Privacy → AI features.

Disabling AI features has no impact on your ability to capture, browse, or search your thoughts.

6. Accuracy and limitations

AI-generated connections and insights are suggestions, not facts. They reflect patterns in language and meaning, not a deep understanding of your life or intentions. The AI may be wrong, irrelevant, or miss important context.

You should treat Untangle outputs as a prompt for your own thinking — not as conclusions. We encourage healthy scepticism. If a connection doesn't resonate, dismiss it.

7. Human oversight

All AI features in Noodle are designed with human oversight built in:

AI outputs are always presented as suggestions that require your active engagement.
No AI feature takes any action on your behalf without your explicit instruction.
You can report unexpected, concerning, or inaccurate AI outputs at ai@noodleapp.xyz.

8. Children and sensitive content

AI features are disabled by default for accounts that indicate the user is under 18. Noodle's AI does not process content flagged as relating to health, finances, or other sensitive categories for the purpose of generating advice — it identifies thematic connections only.

9. Changes to this policy

As AI technology and our use of it evolve, this policy will be updated. We will notify you of material changes by email or in-app notification before they take effect. Significant changes to how your data is processed by AI will always require your renewed consent.

10. Contact

Questions about how Noodle uses AI? Contact us at ai@noodleapp.xyz.