Stand: June 2026
Hermetia combines deterministic calculations from multiple interpretation systems with linguistic processing by AI. This combination can seem very personal because it touches on topics such as self-image, relationships, rhythm of life, resonance and inner patterns. That's exactly why we openly explain where AI is involved, what data it processes, what limits apply and what responsibility remains with humans. The transparency page is part of the basic legal and editorial structure of the website. It supports the requirements of the EU AI Act, in particular the labeling of AI-generated or AI-supported content, and it is intended to give users a clear basis for decision-making before they start a profile, purchase premium content or work with companion, journaling and full profile book.
In Hermetia, AI is primarily used for formulation, condensation and dialogue. These include personal interpretation texts, level texts, daily impulses, explanatory sections in the full profile book, reflection questions, companion answers and editorial variants for different uses. AI helps to formulate calculated patterns in an understandable, calm and coherent manner. AI does not decide which system is right, which statement is true, or which future will occur. It writes from prepared contexts, sets of rules and calculated profile characteristics. If a page, a profile area or a dialog was formulated using AI, this should be recognizable. Hermetia uses AI not as an invisible authority, but as a linguistic tool within clear boundaries.
An important protection mechanism of Hermetia is the separation of three levels: computation, convergence and formulation. The calculation classifies inputs such as dates of birth, system parameters and optional context information according to defined rules from the integrated systems. The convergence engine then checks which motifs recur across multiple systems, which statements should be treated strongly, moderately or cautiously and where contradictions should be made visible. Only then is AI added as a formulation level. She is not given the task of inventing a new truth, but rather of expressing the already structured results in an understandable way. This separation reduces the risk that a language model produces apparently safe statements that cannot be technically or methodically derived from the underlying systems.
The AI model only receives the information that is necessary for the respective content. Depending on the function, these can be calculated profile markers, topic clusters, security levels, shared journal notes, selected companion contexts or already condensed profile sections. Hermetia follows the principle of data minimization: as much context as necessary, as little raw data as possible. The AI should not work unnecessarily with full real names, direct contact details, payment information or freely available raw profiles. Particularly sensitive content such as journal texts or information about relationships are only included in AI-supported functions if they are necessary for the function and covered by the context of use. The goal is a helpful personal dialogue without automatically turning every detail into a training or analysis context.
AI content in Hermetia is not a diagnosis, not a therapy, not a medical advice, not a psychological treatment, not a legal advice, not a financial advice and not a binding life decision. Hermetia does not say whether a relationship should be maintained or ended, whether a person has an illness, whether a career should be chosen, or whether a financial move is right. The content serves for self-reflection. They can provide words for patterns, open questions, and facilitate inner observation. However, they do not replace a qualified specialist or your own responsibility. If content has an emotionally strong impact, is unexpectedly stressful or touches on specific risks, external professional help should take priority.
The Hermetia Companion is planned as a quiet AI companion that responds within your own profile. It is not intended to be a general chat AI, but rather access shared Hermetia contexts: profile book, daily impulses, journaling reflections, relationship interpretations and personal development areas. This allows it to provide answers that are more coherent than a general language model. At the same time, this function needs clear boundaries. The companion should make it transparent from which source an answer is derived, whether a statement is strongly or cautiously worded and when he should not help. Journal texts may not automatically be used without limits. Users must be able to understand what context is used and how to control, delete or not share sensitive content.
AI can make premium content more readable, more personal and more continuous. A full profile book, deeper daily impulses or companion dialogues create value because they translate calculated patterns into longer, differentiated language. Nevertheless, AI must not be used to create artificial pressure, create fear or force users into a paid model. Hermetia is intended to arouse curiosity and invite you to register, but the decision for a paid model must remain understandable. Good conversion comes from clarity, utility, trust and real depth, not from manipulative terseness or seemingly absolute statements. Therefore, purchase arguments must explain what Premium expands: scope, context, history, depth of dialogue, export, comparison and long-term work with the profile.
Hermetia's AI texts must be created in your own language. You may not use any protected works, any copied system descriptions, any long passages of text that are not your own, or any formulations that are problematic under trademark law. If systems of interpretation are mentioned, this is done descriptively, comparatively and with an independent editorial classification. The same principle applies to images, graphics and models: only your own material, licensed assets, self-generated graphics or clearly approved project materials may be used productively. Before the public launch, it should also be checked whether the terms used, product names, system names, image rights and trademark references are legally classified.
AI output is not considered infallible. Prompts, system rules, security boundaries, tonality and content logic must be maintained editorially. Particularly sensitive topics such as health, relationships, crises, grief, sexuality, minors, financial decisions or legal questions require additional barriers and clear information. Before the launch, central AI text patterns, onboarding routes, premium arguments and companion response logic should be checked by a specialist review. This not only affects legal risks, but also tonality, comprehensibility, ethical impact and the question of whether users adequately recognize what is inspiration and what is not.
During the preparation phase, the website intentionally remains on noindex/nofollow because the URL needs to be changed and all content needs to be checked further before the final launch. This setting prevents unfinished content from being built up early in search engines. After the domain move, final editorial approval, legal review and technical quality assurance, the indexing can be opened specifically. AI transparency will grow with the product. As new features such as expanded companion conversations, relationship modules, profile refinement, export or additional EU language versions go live, this page will need to be updated accordingly.