Ethics Principles

Daylogue handles something personal: how you feel, what you think about, and what patterns run through your days. That responsibility shapes every decision we make. These are the principles we build by. They are not aspirational. They describe how the product works right now, and where one of them exists because we got something wrong first, we say so.

Self-awareness, not diagnosis

Daylogue helps you notice patterns in your emotional life. It does not tell you what is wrong with you. It does not diagnose conditions, prescribe actions, or provide clinical guidance of any kind.

When the app surfaces a pattern, it is offering an observation, not a conclusion. "Your stress has been elevated for three weeks" is different from "You have an anxiety disorder." Daylogue does the first. It will never do the second.

The language throughout the product reflects this distinction, and it did not always. Our prompts used to tell the model to "assess the present" and assign you a mood. Arc labels read "recovery" and "decline." An enterprise alert was called burnout risk, which is a clinical concept with an ICD code attached to it.

We went back through the prompts, the alert types, the auto-tags, and the marketing copy and pulled that vocabulary out. The model is now asked to notice what you described, not to assess you. Arcs read "rising" and "dipping." "Burnout risk" became "sustained elevated stress." A rough week is a rough week; it is not a condition, and nothing in the product should quietly hand you one. Patterns are information to sit with, not labels to carry.

Telling you what the architecture actually does

Many apps promise privacy only in their terms of service. We would rather tell you what the system does. Entries are protected in transit by TLS and by per-user access controls in the database. Logging and diagnostic controls are designed to reduce the risk that entry text appears in application logs or error reports; complete coverage is still being verified.

Daylogue's servers read synced entry text because that is how narratives are written and patterns are surfaced. Some legacy account and vault paths contain additional encryption components, but coverage and key custody are still being re-verified. Daylogue does not present them as protection for every entry or as a setting that makes content unreadable to the service. Daylogue is not end-to-end encrypted.

This is the part most privacy pages leave out. A privacy policy can change. A terms-of-service update can quietly expand what a company does with your data. We are not going to tell you the math prevents that here, because it does not. What we can tell you is exactly who can read what, and be held to it.

No manipulation

Daylogue does not use streaks, guilt, or dark patterns to keep you coming back. There are no badges to earn, no counters to protect, no passive-aggressive notifications when you miss a day.

We removed streaks and every remaining gamification element in early 2026. Streaks were already minimal here, and they still had to go. Loss-aversion mechanics are a compulsion risk, and they contradict a product built on no guilt and no shame — a number you are afraid to break is not a reason to write. Missing a Tuesday should cost you nothing, and now it does not.

Transparency about AI

Daylogue uses AI to generate follow-up questions, extract mood and energy signals from your words, identify patterns across entries, and write narrative summaries. This requires your entries to be sent to an AI provider (AWS Bedrock) for processing.

That is a real tradeoff. A journal that reads your patterns back to you has to be able to read what you wrote, which is why Daylogue is not end-to-end encrypted and does not claim to be. AI providers process content under Daylogue's current configuration and written terms. Logging and diagnostic controls reduce risk, but they are not a promise that content can never reach a diagnostic system.

There is also a line we do not cross. Daylogue does not read emotion off your face, and it does not listen for feeling in the sound of your voice. Voice check-ins are transcribed, and it is your words that get read, not how they sounded. Body signals work differently and we would rather be plain about it: heart-rate variability and sleep quality are read alongside the themes in your writing, and Daylogue will tell you when a theme keeps landing on the days your body dipped.

For a detailed explanation of how AI processing works, see How Daylogue Uses AI.

Crisis safety

Daylogue is not monitored for emergencies. Its automated matching may miss crisis content or respond when resources are not needed.

When the system identifies covered crisis language, it is designed to show resources such as the 988 Suicide & Crisis Lifeline and end the automated check-in. That is a fallible resource handoff, not an assessment, prediction, protection, monitoring service, or emergency response.

The voice flow is designed to stop ordinary conversation after resources are shown. Product defects, network failures, unsupported contexts, or missed language can prevent that behavior, so anyone in immediate danger should contact emergency services or a crisis resource directly.

Daylogue is not a crisis service and does not promise to detect or respond to an emergency. Automated resources do not replace professional or emergency help.

An age gate, and knowing who we are talking to

Onboarding asks about your age up front, before you have written anything. You confirm you are 13 or older, or you say you are under 13 — and if you are, that is where it stops, with a message that we will be here when you are ready.

We added that gate because we did not have one. Anyone could open a voice check-in and start talking about their week, and we had no idea whether we were talking with a thirty-year-old or an eighth grader. It is a self-declaration, not an identity check, and we are not going to pretend a checkbox is proof — but a boundary at the front door beats no boundary at all. Onboarding also ends on a box you have to check yourself: that you understand Daylogue is a reflection tool, not a replacement for professional care.

Data ownership

Your data is yours. This is not a marketing line. It describes how the product is built.

• Export. You can use the export tool for the account data and formats it currently supports. The export description identifies current coverage.
• Delete anytime. You can delete your account and all associated data. Deletion is real, not a soft archive.
• We do not sell personal data. Disclosed service providers process data to operate requested features.
• No Daylogue model training. Daylogue does not use your entries to train its own models. Third-party handling follows our current written terms and subprocessor disclosures.

Scope discipline

Daylogue is a wellness tool. It is not a medical device, a clinical resource, or a substitute for professional help. It stays within its lane and clearly communicates what it is and what it is not.

Scope disclaimers appear throughout the product. The AI is prompted to stay within wellness boundaries. The language is deliberately non-clinical. This is not accidental. It is a design principle.

Scope creep in wellness apps is dangerous. A tool that gradually starts acting like a therapist puts users at risk by creating a false sense of clinical support. Daylogue resists that drift by building scope boundaries into the product itself, not relying on users to understand the limitations on their own.