Who would you like
alongside you?
Choose someone who fits, shape a companion from a blank beginning, or return to one you already know.
Explore AETHERA’s five realms → · An interactive world concept.
Our principles and current limitsAlongkind’s public commitmentOur requirements for permission, bounded AI authority, Stop and Revoke, evidence, and final human control.
Read our public commitmentHuman control: requirements and tested scope
- Human authority is final.
- Meaningful action must require permission.
- Unclear authority must prevent action.
- In the username receipt beta, Stop/Revoke blocks execution when checked before the transaction commits.
- Admitted requests in that beta produce signed attempt and outcome receipts. This is not coverage of every app action.
Acceptance requirement — implementation evidence neededUnknown or ambiguous authority → no direct or indirect action; permission withdrawal → immediate safe halt and closure of every derived route; every attempted, blocked or completed action → a durable, independently verifiable receipt the agent cannot alter or erase.
Live receipt evidence covers the separately deployed username reservation beta only. It does not demonstrate complete Stop/Revoke propagation to external tools or delegated agents, or protection against privileged host administrators. The browser test on the results page is a local demonstration. Life-stage categories are catalogue filters, not parent PIN locks or age verification. Young Explorer is a preview, not approved for unsupervised child use.