Open source and access

Citrate is built to be open. The entire chain and its application layer are open sourced before mainnet, in January 2027. Until then, access to the core is deliberate, and that is a security decision, not secrecy.

How we work

We research and build in private, and we open source a component once it has findings worth publishing and third-party validation behind it, from auditors or independent researchers. Publishing code the day it is written, before it is audited and before the design has settled, invites two failures we refuse: IP theft during the window when a design is most copyable, and a wider attack surface on infrastructure that is still moving. Opening a component after it is validated is the stronger position, for the network and for the people who build on it.

What is open today

Some work is public now, because it is meant to be read and reproduced:

• NAT is the model architecture. Memory-safe Rust, formally specified, public. You can read the source and reproduce the results.
• American Learning Federation (ALF) is the cooperative that trains NAT through federated learning.
• agentile-skills is the engineering methodology, installable by anyone.

Getting access to the rest

If you want to build on the parts that are not yet public, ask. Request access on the website or by email. Once your application is processed, approved contributors receive privileged access to every repository in the federation, with one exception: the private repositories of clients and employees, which stay closed for their sake, not ours.

• Request access: citrate.ai/contact, or email hello@citrate.ai.
• Already building: NAT and ALF are open now. Start there.

The commitment

The whole chain and the whole application layer are open sourced before mainnet, in January 2027. That is a date, not an aspiration. What is private today is private on a schedule we set, for reasons we can name, and every part of it has a public release ahead of it.