Show HN: SQLite Graph – Graph database with Cypher queries (alpha, but working) (github.com)

AgentFlare AI announced SQLite Graph, an alpha SQLite extension that turns SQLite into a lightweight graph database with end-to-end Cypher support. The v0.1.0-alpha.0 release lets you CREATE, MATCH, FILTER (WHERE) and RETURN graph patterns using Cypher, or manipulate the graph via SQL functions (graph_node_add, graph_edge_add, graph_count_nodes, etc.). It includes a virtual-table interface, Python 3.6+ bindings, basic graph algorithms (connectivity, density, degree centrality), and security/thread-safety fixes — useful for rapid prototyping, local agents, or embedding graph capabilities into existing SQLite workflows. Technically the extension implements a full Cypher execution pipeline (lexer/parser → logical planner → physical planner → volcano-style executor) and claims 70/70 CREATE TCK tests passing for CREATE semantics. Performance numbers are strong in microbenchmarks (≈300k nodes/sec, ≈390k edges/sec) though longer scalability tests are currently limited (tested up to ~1k nodes/edges, persistence via SQLite files). Advanced features — variable-length paths, bidirectional matching, aggregations, full openCypher compliance, query optimization and advanced algorithms (PageRank, betweenness, community detection) — are on the roadmap for v0.2–v1.0. This alpha is ready for testing and feedback but not production; source, builds and examples (libgraph.so, python demos, ROADMAP.md) are available on GitHub.