NexusFS
A verifiable offline-first distributed filesystem. NexusFS records writes as signed operations, derives a deterministic CRDT view, and uses Merkle proofs to support third-party verification.
Storage whose history can be checked
Every accepted change becomes a signed operation in an append-only log. Replicas derive the same filesystem view through deterministic CRDT projection, while a Merkle accumulator supports inclusion checks and replay.
QUIC provides resumable peer transport for environments where devices disconnect, move between networks, or operate behind NAT.
- 01Signed, monotonic filesystem operations
- 02Deterministic conflict-free projection across replicas
- 03Merkle inclusion evidence and independent replay
- 04Offline-first transport and content-addressed blobs
Connecting replication, provenance, and audit
The research question is not only whether replicas converge, but whether another party can reconstruct why a specific state exists. Signed operations and deterministic replay make that history inspectable.
NexusFS is also intended as a storage substrate for evidence produced by Wardex and SafeDrop, where a receipt is useful only if its later custody can be checked.
Evidence boundary
NexusFS is explicitly a prototype. The verifier CLI, broader failure testing, and production hardening remain active work; the page does not present the design as production-proven.