NamiDB Documentation

NamiDB is a graph database engine. Its state — manifest, WAL, SSTs, schema — lives as plain objects in an S3-compatible bucket. The engine ships in three shapes (embedded library, HTTP daemon, managed cloud); all three speak the same Cypher and write to the same bucket layout.

These docs cover the engine end-to-end. If you’ve never used NamiDB before, the 30-second quickstart is the fastest taste; the S3 walkthrough shows the headline use case.

Documentation map

Get started

What NamiDB is, the three deployment shapes, how to install each client, and the two quickstart paths.

• What is NamiDB
• 30-second quickstart
• Your graph in S3
• Choose a deployment
• Install

Concepts

The mental models behind the engine — why object storage as source of truth works, what a snapshot is, how single-writer-per-namespace is enforced, where the caches live.

• The bucket is the database
• Three deployments, one engine
• Namespaces & multi-tenancy
• Snapshots & epoch fencing
• LSM on object storage
• Caches

Cypher reference

The exact GQL (ISO/IEC 39075:2024) + openCypher 9 subset NamiDB parses, plans, and executes today. Includes the 12 in-scope LDBC SNB Interactive queries.

• Supported subset
• Read queries
• Write queries
• Functions & operators
• LDBC SNB IC01–IC12

SDK reference

One engine, four surfaces. Same Cypher across all.

• Python — pip install namidb
• Rust (embedded) — cargo add namidb
• CLI — namidb parse / explain / run
• HTTP API — namidb-server REST endpoints

Operations

Environment variables, the URI grammar, all six storage backends, the Docker Compose self-host recipe, observability, backups, and tuning knobs.

• Configuration · URI grammar
• Storage backends: AWS S3 · R2 · GCS · Azure · MinIO/Tigris/LocalStack · Local
• Self-host with Docker Compose
• Observability · Backup & restore · Tuning

Cloud

Managed multi-tenant SaaS on namidb.com, per-namespace scale-to-zero, closed beta.

• What is NamiDB Cloud
• Request access
• Cloud vs self-hosted

Internals (RFCs)

18 design RFCs covering the storage engine, SST format, query engine, cost-based optimizer, factorization, and caches. The canonical source of “why the engine looks like it does.”

• RFC index
• Storage engine: RFC-001 · RFC-002 · RFC-003
• Query language: RFC-004 · RFC-008 · RFC-009
• Optimizer: RFC-010 → 016
• Executor & caches: RFC-017 · RFC-018 · RFC-019 · RFC-020

Community

How to engage with NamiDB development, the RFC process, security reporting, license terms.

• Contributing
• RFC process
• Security
• License (BSL 1.1)

Changelog

All notable changes across the engine, Python bindings, server, and CLI.

Common entry points

• I want to write Cypher right now → 30-second quickstart
• I want to use my own S3 bucket → Your graph in S3
• I want a REST endpoint to call from my app → HTTP API + Self-host recipe
• I want to know what Cypher works → Supported subset
• I want to read how the storage engine works → RFC-001 → RFC-002
• I want to migrate from Kùzu / Neo4j → Cypher reference and open an issue if anything is missing
• I want the managed product → NamiDB Cloud

Project info

Resource	Where
Engine repo	github.com/namidb/namidb
PyPI package	pypi.org/project/namidb
Issues & RFCs	github.com/namidb/namidb/issues
Security	security@namidb.com
General	hello@namidb.com
Website	namidb.com
License	BSL 1.1 → Apache 2.0 after 3y