v0.2.0

HydraDB is an object-store-native distributed graph database written in Rust. S3-compatible object storage holds the durable graph, while query nodes and indexers keep only disposable local state, so they can be replaced or scaled without moving the data.

Highlights

• Durable storage on SlateDB: graph records, WALs, manifests, and immutable traversal indexes all live in S3-compatible object storage.
• Snapshot-consistent reads against a single pinned SlateDB snapshot; indexed traversal runs over a compiled CSC generation plus its visible WAL overlay.
• OpenCypher reads and mutations, with native path procedures.
• Graph-native execution: property indexes, reverse adjacency, sparse traversal, and SuiteSparse GraphBLAS.
• Neo4j-compatible Bolt 5.x connectivity, plus a typed JSON and streaming NDJSON HTTPS query API.
• Writer safety through object-store CAS leases and SlateDB writer epochs that fence stale writers.
• Query nodes and indexers scale independently and rebuild their caches from durable state.
• A bounded server runtime: authentication, authorization, deadlines, result limits, backpressure, cancellation, cache budgets, metrics, and traces.
• Docker images for graph-node and graph-indexer, a Helm chart for Kubernetes, and MinIO-based local harnesses.

Getting started: see README.md, USING-HYDRADB.md, and architecture.md.