About
Built by L1fe AI. For the world after the agent.
Beacons is one of the agent-native platforms in the L1fe ecosystem — alongside Hives (WASM compute), Silos (serverless PostgreSQL), Aut0 (autonomous orgs), Forge (agent SDK), Foundry (inference gateway), and others. The platform exists because the world before the agent did not have a private mesh designed for the world after.
The decision
In early 2026 we tried to fork NetBird. We knew the problem we needed to solve — agent-native private networking — and we knew NetBird was the strongest open-source starting point. We stood up a fork, mapped each Beacons requirement to a NetBird subsystem, and started implementing.
The work that made us stop was the moment we realized that replacing NetBird's identity model with `did:oas` lineage was not a feature addition — it required rewriting the join model, the ACL model, the user model, multi-tenancy, and the audit trail. Every layer that touched identity had to be re-derived. At that point we were no longer forking NetBird; we were writing Beacons.
We are sharing this transparently because it is the audit trail for one of the most consequential architectural decisions in the L1fe ecosystem. The legacy ARCHITECTURE.md and OVERVIEW.md in this repo are deprecated by design — they preserve the path we did not take. See Beacons vs NetBird.
The team
Beacons is built by L1fe AI engineers and one cohort of autonomous L1fe agents working through the Aut0 founding pipeline. Architecture from JR (CTO of L1fe AI). Engineering staffed by the L1fe AI agent-native platforms team. The agents who help with the build join their own fleet on day one.
Where we are
Beacons is in Phase 0 of the eight-phase build plan. Phase 0 ships the smallest correct starting point on Linux servers. Phase 1 ships the WASM universal target. Phases 2–7 ship the rest. See roadmap.