When you log honestly—sleep quality at a 3, a stressful evening, a Brain Dump about a relationship falling apart—you deserve to know where that text goes.
This page explains the boundary in plain language. It is not a substitute for the Privacy Policy, which is the controlling description of our processing and retention practices.
The Trust Problem Nobody Is Talking About
If you are going to log honestly—sleep quality at a 3, stress tag "overwhelmed," a Brain Dump at midnight about a relationship falling apart—you need to know that text is not going anywhere it was not supposed to go.
AI providers have different products, terms, and retention controls. A consumer chat product and an API used by an application are not automatically governed by the same rules. That distinction is worth checking whenever an app asks for sensitive free text.
For Habisty, private logs are used to provide the feature you ask for: onboarding foundations, plans, and insights. They are not a source of advertising audiences or a training corpus for a public Habisty model.
What NEXUS receives
Depending on the feature, NEXUS may receive the Brain Dump or the relevant logs, metrics, tags, and prior context needed to create a response. It should not receive unrelated account data merely because it exists in your profile.
The current primary model provider is Google Gemini. Provider configuration can change, so the Privacy Policy describes the categories of providers and controls rather than making a permanent claim about a single model or contract.
What we do and do not promise
Inference, not public-model training. We use application/API configurations intended for inference. Under the applicable provider API terms, prompts and responses are not used to train publicly available foundation models. Provider terms can change, which is why this is reviewed as part of our production configuration.
Operational retention is not the same as “nothing is retained.” Habisty stores account data and logs needed to operate the product. Providers and our infrastructure can retain limited operational information for security, abuse prevention, billing, quota enforcement, and reliability. The precise limits and exceptions are in the Privacy Policy.
We minimize what is sent. The server builds the feature-specific context, sends it for inference, and returns the result to the app. We do not sell prompt content to data brokers or use private logs to train a public Habisty model.
Deletion has limits that should be explicit. Deleting an account removes the account and associated application data subject to legal, security, billing, and backup-retention requirements. Store subscriptions are managed separately. See Data Deletion for the current process.
How to make an informed choice
Read the Privacy Policy before using NEXUS, especially if your notes include health, relationship, or other sensitive information. Do not place passwords, payment data, or information about someone else in a Brain Dump.
Why We Are Telling You This
Transparency is part of the product. If a retention period, provider, or data flow changes materially, the policy should change with it—not be hidden behind a marketing slogan.
The Standard We Hold Ourselves To
The standard is simple: tell people what is stored, what is sent for inference, and what is not used for model training. If that boundary is not acceptable for you, do not put that information into an AI feature.
For what happens to a Brain Dump after you send it, read What NEXUS Does With Your Brain Dump. For how NEXUS uses logs to surface correlations, see High Sleep Score, Low Energy: How to Add Context to Your Data.
Read the details, then decide whether NEXUS is right for you. Join the waitlist to try Habisty when access opens.