What we owe you

Fatement handles something fragile: the moment when someone is honest about being stuck. That honesty deserves protection beyond what privacy law requires. These are our commitments.

Your data is not a product

We do not sell or license your interview data. Your archetype results are not shared with third parties. When you delete your data, it is gone from our database and from every provider that acts on a deletion request from us.

There is one part of that we cannot promise, so we will not. Text you chose to send to an AI provider has already arrived there, and what Anthropic, Google or TypeSafe keeps of it afterwards is set by their terms rather than by ours. The privacy policy says what we know about each. The delete button does not reach it, and no button we could build would. We would rather you knew the edge of the promise than found it later.

We may use anonymized, aggregated analytics to understand how the tool is used and to inform marketing, but only with your consent and never in a way that identifies you personally.

Your data does not train AI

Your interview responses are not used to train AI models. That covers both kinds of input: the text you write, which Anthropic processes, and the audio you record if you use voice input, which Google transcribes. Both are on paid API terms that exclude inputs from model training and from product improvement. The free tiers of these services do not exclude either, which is a reason we are not on them.

A third provider, TypeSafe, reads the answers on your Life Audit sheet when you ask that page for a sharper read of its panels. Nothing goes to them unless you ask for that read. We do not yet have the same training exclusion from them that we have from Anthropic and Google, so the promise in this section does not cover that one request. It is named here because you should be able to read where your writing can go before it can go there, not after.

Transparency over performance

When the AI fails or produces low-confidence results, we tell you. Fallback results are clearly marked. We would rather give you an honest “approximate result” than a polished lie.

Our methodology page explains how the engine works, including its limitations.

Autonomy over engagement

We do not optimize for time-on-site, session count, or return visits. We do not send push notifications, retention emails, or re-engagement campaigns. We do not have your email.

The best outcome is that you use Fatement once, get clarity, and move on with your life. If you come back, it should be because you found it useful, not because we engineered a habit loop.

Safety without gatekeeping

When our safety system detects distress, it surfaces resources. It does not block the session, lock your results, or force you into a workflow.

We believe in your capacity to make decisions about what you need. Our job is to provide information and options, not to control your experience.

Anti-validation

Fatement is deliberately built to avoid the validation trap. The engine will not tell you what you want to hear. It will not praise your answers. It will not use phrases like “you’re doing great” or “that must be really hard for you.”

This is not coldness. It is respect. You came here because surface-level comfort wasn’t working. We honor that by taking your situation seriously enough to push you on it.

No dark patterns

We do not use:

• Countdown timers or artificial urgency
• Shame-based messaging to drive conversions
• Hidden costs or unexpected charges
• Confusing opt-out flows
• Pre-checked consent boxes
• Misleading claims about what the tool can do

Accountability

These commitments are not marketing copy. They are operational constraints we build against. If you believe we have failed to uphold any of these principles, we want to hear about it: hello@fatement.net.

See also: our complaints process.