Solo dev, Montenegro. Building an open protocol for cryptographically validating digital work — designed to run on a $30 phone offline or across interplanetary delays.
What exists and runs today:
- Layer 1 (Python) — Local validation with CRYSTALS-Dilithium3 + SPHINCS+ (FIPS 204/205). SHA3-256 hashing, SQLite-backed DAG. 50 tests. Identity generation in 2ms, file validation
in 1-2ms. MIT licensed. https://github.com/navigatorbuilds/elara-layer1
- Layer 1.5 (Rust) — DAM Virtual Machine with PyO3 bindings. Same crypto via liboqs, byte-identical wire format to Python. 5-tuple dimensional addressing, tiled storage, in-memory
DAG index, 9 DAM operations, parallel batch verify via rayon. 72 Rust + 21 cross-language tests. Wire format serialization in 292ns. https://github.com/navigatorbuilds/elara-runtime
- Layer 3 (Python, MCP server) — Persistent cognitive architecture for AI assistants. Memory, emotional modeling, episodic recall, autonomous reasoning, dream processing. 39 tools,
~26K lines. On PyPI: pip install elara-core. https://github.com/navigatorbuilds/elara-core
What's specified but not built:
- Layer 2 (network consensus via Adaptive Witness Consensus)
- Native hardware (custom silicon where the 5D mesh maps to physical substrates — PCM for time, memristive crossbar for concurrency, photonic mesh for zones). This is a long-term
vision, not a promise. The protocol runs indefinitely on commodity hardware.
The data structure is a Directed Acyclic Mesh (DAM) — a DAG extended with 5 dimensions: time, concurrency, zone topology, classification projection, and AI analysis. No blockchain.
No universal clock. The minimum viable network is one device.
Crypto: Post-quantum from genesis. Dilithium3 + SPHINCS+ dual signatures. No classical crypto to migrate away from later.
Honest limitations: Layer 2 consensus is specified but unimplemented. The formal Byzantine proof is pending. ZKP layer uses classical elliptic curves in Phase 1 (migration path to
lattice-based ZKPs defined but years away). Hardware whitepaper is speculative engineering — none of it is needed for the protocol to work.
US provisional patent filed (63/983,064). Whitepapers OTS-timestamped on Bitcoin.
Happy to answer technical questions about the DAM structure, the PQC choices, or why I think the mesh model works better than a chain for this problem.