# AgencyDomains

> The formal architecture of the Agentive World. An open specification for building agentic and
> agentive AI in production: the AgencyDomain as the minimal unit of deployment; four layers
> (Interaction, Cognition, Autonomy, Access); canonical primitives — Botlet, Capability,
> Trust Infrastructure, Facet. Edited by GegoLabs. Pre-1.0.

## The book

- [AgencyDomains — canonical manifesto for agents (English, Markdown)](https://agencydomains.org/agents.md): the book distilled for an agent to reason within the framework. v0.4.
- [AgencyDomains — same manifesto as plain text (.txt, alias)](https://agencydomains.org/llms-full.txt)
- [AgencyDomains — web edition (English)](https://agencydomains.org/agencydomains/): the full book, navigable. (Figures are still in Spanish; their English versions are forthcoming.)
- [AgencyDomains — PDF (English)](https://agencydomains.org/agencydomains/AgencyDomains-v0.4.pdf)
- [AgencyDomains — Spanish edition (web)](https://agencydomains.org/es/agencydomains/) · [PDF](https://agencydomains.org/es/agencydomains/AgencyDomains-v0.4.pdf) · [agents manifest](https://agencydomains.org/es/agentes.md)

## Reference implementation

- [Vergis (code, AGPL-3.0)](https://github.com/gegolabs/vergis): the AgencyDomain made operational — Botler + Mira.

## Coming soon

- Postchat · The Agentive World (the destination)
- Aura · The Agentive Path (the route)