Roadmap

NOEMA ships as a family of connected products, built in order. Each stage delivers something you can use on its own, and each becomes the foundation for the next.

Phase 1: Noema Client

The flagship app. Chat with open models running fully in your browser (WebGPU) or on your own hardware via Ollama, private by default because the work never leaves your machine. In beta now, updated continuously.

Phase 2: Noema Engine

The local-first router. It evaluates every workload and decides where it runs: your silicon by default, a frontier API only when you explicitly allow it, and eventually the sovereign network. This is the layer that enforces "intelligence stays local until it must scale."

Phase 3: Noema Network

The sovereign compute layer. Independent GPU operators run workloads inside hardware enclaves (TEEs), reached through onion-routed circuits, secured by $NOEMA staking and slashing on Base. Includes the Provider Node CLI and, once battle-tested, the public SDK for developers.

Timelines will be published as each phase approaches. Order is the commitment; dates follow.