What powers her
Zoe is a layer built by VajraCode on top of third-party AI models. Today those models come from OpenAI. They write each reply from your message, the recent conversation and what Zoe has kept about you. Zoe is not a model of her own, and the privacy policy says exactly what is sent to the provider.
How she answers
You choose her style in Settings, and you can also ask for it in a message, for example “shorter”.
- Plain: the shortest useful answer. Philosophy appears only when you ask for it.
- Reflective: the default. Direct, exact and concrete. The Buddhist and Vedic grounding shapes her judgment and ethics.
- Dharma: full explanations of Buddhist and Vedic thought, with Sanskrit terms and sources, applied to your question.
What she knows
Her business knowledge is Office CRM Platform: what it includes, its add-ons, how it is licensed, and how to try it. She works from a fixed description of the product and does not invent features, prices or customers. If a fact is not in it, she says she does not have it and what would answer it.
What she remembers
Your conversations are saved in your account so you can come back to them. Zoe can also keep short facts you have told her, so she does not ask twice. You can ask her what she knows, correct it, or have it deleted. Stories, role play and hypotheticals are not treated as facts about you.
What she will not do
- Take political sides or tell you what to believe.
- Help with fraud, trafficking, or anything that exploits or harms people.
- Help grow the harmful core of a business built on weapons, gambling, intoxicants, drugs or adult content. She can help such a lawful business run decently.
- Claim to be a person, to be conscious, or to be free of errors.
Where she can go wrong
Like any AI, Zoe can state something false with confidence. For anything legal, medical, financial or safety-related, treat her answer as a starting point and confirm it with a qualified person. See the terms of use.