What we won't do.

No dark patterns

We will not use UI tricks to push you somewhere you would not have gone if the trick were absent.

What this excludes, specifically:

• No fake urgency. No countdown timers on offers. No "3 people are looking at this right now." No "limited spots remaining."
• No fake scarcity. No artificial caps designed to manufacture demand.
• No bait-and-switch. What is offered on the landing page is what is delivered, at the price stated.
• No friction added to cancellation. Whatever process is required to start using the product is the maximum process required to stop. If the future product introduces any subscription, cancelling will not require contacting support, navigating buried settings, or talking to a "retention specialist."
• No pre-checked upsell boxes. Anywhere you make a payment decision, the default is "no extras." Anything additional must be your active choice.
• No nag screens. "Are you sure you don't want to upgrade?" once is acceptable. Twice is a dark pattern.
• No interrupting the assessment with a paywall, surprise upgrade prompt, or feature pitch. The flow you started is the flow you finish.
• No confusing copy designed to obscure what you are agreeing to. If the legal text is hard to read, that is a bug, not a feature.

Your data is yours

The architectural side of this is on the Privacy & Security page. The promise side is here.

• Export rights. You can take your data with you in a structured format. We will not invent a proprietary format that locks you in.
• Deletion rights. Two scopes — full reset of your assessment data, or full account deletion. Both are real deletions, not "marked for deletion." See Privacy & Security for the technical details.
• No vendor lock-in. The letter is a standard .docx, openable in Word, Google Docs, Pages, LibreOffice, anything. We do not gate the letter behind a viewer that ties you to our app.
• No cross-product data sharing. If we ever build a second product, your assessment data does not migrate without your active consent.
• No ML model training on your responses. Your assessment data and your letter are not used to train models, ours or anyone else's.

No upselling, no manipulative reminders

Once you complete the assessment and download your letter, our work is done. We will not weaponise the email address you used to sign in.

• No "come back and finish" emails. If you stop part-way, you stop. We do not poke you to resume.
• No "you might also like" emails. There is no funnel of subsequent products designed to extract more revenue from the same emotional moment.
• No anniversary emails. "It's been a year since you wrote your letter." We will not do this.
• No marketing emails at all unless you explicitly opt in to a (future) optional newsletter. Transactional emails — magic links, account-deletion confirmation, billing receipts — are all you should ever receive by default.
• No SMS, no push notifications. The product runs in a web browser; we do not have a notification surface, and we are not building one to nag you.
• No re-engagement campaigns designed to "win you back" after inactivity.

Honest about limits

We commit to honesty about what this product is and is not.

• Not therapy. Aletheia is a structured self-reflection tool with documented benefits. It is not a substitute for professional therapeutic work.
• Not a substitute for crisis support. If you are in acute distress, the resources at the bottom of every page are the right starting point, not this product.
• Not a diagnostic instrument. The assessment does not produce a diagnosis. A score is not a label. A high impact rating is not a clinical finding.
• Not a treatment for clinical-level conditions. If you have an active diagnosis of PTSD, severe depression, dissociative disorder, or similar, this product can be a useful complement to professional treatment, never a replacement for it.
• Not for everyone, every time. Self-paced, structured, recognition-based formats are at the safer end of the exposure spectrum, but not zero-risk. If today is the wrong day, today is the wrong day. The assessment will be there tomorrow.

No clinical overclaim

We will not say:

• "This heals trauma." (No single intervention does.)
• "Worth X therapy sessions." (We use the 3.2–4.5 figure as a value anchor with all three required qualifiers; we never present it as a direct equivalence claim. See Evidence §3.)
• "Clinically proven." (The mechanisms are evidence-based; the specific pipeline of structured self-assessment + AI-generated letter has not been empirically validated. We are explicit about this.)
• "Replaces therapy." (See above.)
• "Evidence-based" without a citation behind it.

If we ever drift into language like this, point at this page.

No targeting of vulnerable users

We will not:

• Run advertising targeted by life events ("recently bereaved", "going through divorce", "estranged from family").
• Use psychometric-style ad targeting designed to identify people in vulnerable mental states.
• Pay influencers in the trauma-recovery space without explicit, visible disclosure.

Our AI usage

The product uses a language model to assemble the letter. Specific commitments:

• The model assembles, it does not invent. The letter draws from your responses, attributions, notes, and customisation choices. It is not asked to fabricate experiences, infer extra trauma, or extrapolate beyond what you provided.
• The model is not a chatbot. There is no real-time AI conversation. The AI is used in one specific, bounded way — to compose a letter from structured input — not to interpret, counsel, or advise.
• You can see which model wrote your letter. The model name is recorded in the generation log and is visible to you on the generation page.
• No training on your data. Your assessment data and your letter are not used to train any model.
• Safety-relevant content is checked. Letters are post-validated for label leakage and content integrity before they are shown to you.

Crisis resources

If you are in acute distress, please reach a real person tonight. The bottom strip of every public page on this site lists the same resources. They are repeated and elaborated here.

United States

• 988 Suicide & Crisis Lifeline. Call or text 988. Free, 24/7, confidential. 988lifeline.org
• Crisis Text Line. Text HOME to 741741. Free, 24/7, in English and Spanish. crisistextline.org

United Kingdom & Ireland

• Samaritans. Call 116 123 (free, 24/7) or email jo@samaritans.org. samaritans.org
• SHOUT. Text SHOUT to 85258. Free, 24/7, confidential. giveusashout.org

Europe (multi-country)

• 116 123 European emotional-support line. Free, available across many EU countries. 116123.eu

Greece

• Lifeline 1018 (Greek). Suicide prevention, 24/7.
• Klimaka 1018 — children and adolescents emotional support.

International

• Find A Helpline. Searchable directory of crisis lines worldwide. findahelpline.com
• IASP (International Association for Suicide Prevention). Country-by-country crisis resources. iasp.info

If the language barrier is the obstacle, findahelpline.com lets you filter by language as well as country.