Argorix Lang is open and early — built for people thinking deeply about AI safety, agents, compilers, runtime governance and verifiable execution.
Pick a surface and open a pull request. Every layer is auditable on purpose.
Open an issue, propose a design, or join a discussion on the language and runtime.
Contribute to the parser, IR, bytecode VM, or the sandboxed adapter framework.
Strengthen the public conformance suite and contribute threat models.
Write policy packs that map to EU AI Act, ISO/IEC 42001 and local regimes.
Argorix Lang follows the Internet of AI Agents research and builds its identity layer on two Chilean researchers’ protocols.
The research line guiding how sovereign agents discover, identify and trust each other.
projectnanda.org ↗The trust-and-evidence model behind Argorix’s identity, credentials and trust ledger.
Edison Vásquez Droguett · ChileThe citizenship-and-jurisdiction framing behind binding a passport to a country and a legal regime.
Danilo Naranjo Emparanza · ChileApache-2.0, built in Rust, designed to be inspected, tested and governed in the open.
When in doubt, the runtime stops. No capability ships without governance and evidence.
Identity, jurisdiction and evidence are first-class — so control stays in public hands.