WHY NOW
Models keep advancing.Memory shouldn’t reset.
We maintain an independent memory layer outside the model, so that accumulated context is not lost when models change.
Today, most personal AI memory is confined to a single application or platform. Switching applications or providers often means accumulated context cannot be carried over, and a new model generation may change how the AI expresses itself and responds. Aetern AI builds independent memory and identity infrastructure outside any single model or application, so that what accumulates through long-term interaction remains traceable, correctable and continuous.
PRINCIPLES
- I
Shaped byevery interaction.
The value of a personal AI accrues through long-term interaction: its understanding of the user, its own character and the history they share all take shape over time.
- II
Models change.Identity should endure.
A model determines how an AI reasons. Its identity and its understanding of the user are shaped by interaction, and should persist independently of any model — across models, applications and devices.
- III
Memory belongs to the user.Under the user’s control.
All memory formed between a user and an AI belongs to that user: it can be viewed, corrected, deleted and carried elsewhere. We do not train models on personal data.