Welcome to the NodeDB docs. These guides explain what each engine does, when to use it, and how to get started. For full API reference, see the cargo docs or the API reference (coming soon).
- Getting Started — Prerequisites, build, run, first queries
- Architecture — How the three-plane execution model works (now includes cross-engine identity)
- Query Language — Full SQL reference (DDL, DML, engine-specific syntax, functions)
- Protocols — Six wire protocols (pgwire, NDB, HTTP, RESP, ILP, Sync)
- Vector Search — HNSW + Vamana, frontier quantization, adaptive filtering, vector-primary collections, hybrid search
- Graph — CSR adjacency, 13 algorithms, MATCH patterns, GraphRAG
- Documents — Schemaless (MessagePack + CRDT) and Strict (Binary Tuples, OLTP)
- Columnar — Per-column compression, predicate pushdown, HTAP bridge
- Timeseries — ILP ingest, continuous aggregation, PromQL, approximate aggregation
- Spatial — R*-tree, geohash, H3, OGC predicates, hybrid spatial-vector
- Key-Value — O(1) lookups, TTL, atomic INCR/CAS, sorted indexes (leaderboards), rate gates, SQL-queryable
- Full-Text Search — BM25, stemming, fuzzy, hybrid vector fusion
- Array (NDArray) — Multi-dimensional tiles with Z-order indexing, compression, bitemporal support
- AI Pattern Guides — Index of all AI/ML guides
- RAG Pipelines — Basic, hybrid, filtered, parent-document, conversational RAG
- GraphRAG — Entity extraction, graph expansion, community summarization
- Agent Memory — Episodic, semantic, working memory with scheduled consolidation
- On-Device AI — NodeDB-Lite for offline RAG, CRDT sync, WASM, privacy
- Multi-Modal Search — Multiple vector columns, cross-modal CLIP, RRF fusion
- Feature Store — Training features, point-in-time lookups, batch export, online serving
- CDC for Inference Triggers — Embedding pipelines, graph re-indexing, model output routing
- Evaluation Tracking — Experiment metrics, retriever comparison, drift detection
- Multi-Tenancy for AI SaaS — WAL-level isolation, RLS during search, per-tenant budgets
- Bitemporal Queries — System time and valid time, audit trails, corrections, compliance
- Cross-Engine Identity — Surrogate bitmaps for fused multi-engine queries
- WASM (Experimental) — Browser and Node.js build, offline-first CRDT sync
- NodeDB-Lite — Embedded database for phones, browsers, desktops (separate repo)
- Security — Overview, encryption, quick reference
- Authentication — Users, passwords, API keys, JWKS, mTLS
- Roles & Permissions — RBAC, GRANT, REVOKE
- Row-Level Security — Per-row filtering with
$auth.*context - Audit & Change Tracking — Hash-chained audit log, SIEM export,
updated_atpatterns - Multi-Tenancy — Isolation, quotas, backup, GDPR purge
- Real-Time — LIVE SELECT, CDC change streams, consumer groups, webhook delivery, durable topics, cron scheduler
- CLI (
ndb) — Terminal client usage and configuration (separate repo) - Studio — GUI client (Dioxus desktop + web) (separate repo)