Byzantine fault tolerance, deployed where your data is governed.
WeilChain is the DAG-based replicated state machine under the Weilliptic stack. Each WeilPod is an independent chain in the region or organization you choose, and every write is ordered by consensus before it commits.
No single operator is trusted.
Each region runs its own client endpoints, Sentinel cluster, WeilPod and IdemPod. Requests are validated and authenticated at the Sentinels, then ordered by the pod's validators before they commit. The Senate provisions pods and coordinates across them.
What belongs on WeilChain, and what does not.
BFT ordering buys a stronger fault model at the cost of throughput. That trade is right for facts that must survive a compromised operator. It is wrong for high-volume data movement, so the platform does not put the data plane here.
| Workload | Placement | Layer | Why |
|---|---|---|---|
| Agent identity and grants | WeilChain | Autos | Must hold when a host or operator is compromised |
| Authority and policy decisions | WeilChain | Writ | Ordered and attributable, so any decision can be replayed |
| Shared agent memory | WeilChain | Pensieve | State shared across agents with no trusted host |
| Evidence of execution | WeilChain | Receipts | Ordered before it is written, independently verifiable after |
| High-volume event and data streams | Data plane | WeilStream | Throughput-bound. Consensus ordering adds cost without adding safety |
The full specification is in the docs.
Architecture, applet anatomy, state semantics, the WIDL reference and the CLI. Deploy a local platform for evaluation in one sitting.
Open WeilChain docs