ModelChain

Append-only blockchain in C++20 for versioning ML models and datasets. The chain stores cryptographic hashes and metadata of ML artifacts; large weights live off-chain in a content-addressable store. Target: a trusted 3–5 node cluster that synchronizes over a coroutine-based P2P protocol.

Prerequisites (Windows 11)

• Visual Studio 2022 with the "Desktop development with C++" workload
• CMake 3.28 or newer
• Git

Bootstrap

Install a local, pinned vcpkg instance and pin the dependency baseline:

./scripts/bootstrap.ps1 -VcpkgRef 2025.06.13

Omit -VcpkgRef to use the vcpkg default branch (not reproducible).

Build

./scripts/build.ps1 -Config Release

Test

./scripts/test.ps1 -Config Debug

Layout

Each library under libs/<module> is self-contained (include/, src/, CMakeLists.txt) with a single responsibility and an explicit public API:

• serialization — zero-copy binary (de)serialization over std::span
• crypto — SHA-256 hashing and the Hashable concept
• storage — content-addressable store for off-chain artifacts
• ledger — blocks, append-only Merkle tree, chain validation
• validator — streaming chain validation via std::ranges
• registry — MLflow-like model/version/lineage API
• p2p — coroutine-based peer discovery and block synchronization
• observability — structured logging and metrics

See docs/architecture.md for the single-schema diagram and docs/acceptance-criteria.md for non-functional requirements.