THE FRONTIER DATABASE

Graphs, tables and vectors in one server.

Store data, the relationships between it and the rows that belong to it. Query with Cypher and SQL, search by meaning, explore in the browser and run graph algorithms, all on your own infrastructure.

Cypher

Create nodes, traverse relationships and aggregate results in standard Cypher, over the Bolt drivers you already use.

Query guide

Tables and SQL

A node holds typed tables with keys, indexes and constraints. SQL with joins, window functions and recursive CTEs reads and writes them in the same transaction as the graph, with results checked against SQLite.

Tables and SQL

Open source drivers

C#, Rust, Java, Go, Node, C++ and Python, under Apache-2.0. Parameterised queries, native types and spatial values, with no third-party runtime packages.

Drivers

Indexes

Property lookups, full-text search, nearby points, intersecting shapes and vector similarity. Hash, range, composite and element indexes cover the rest.

Indexes

Graph algorithms

Find communities, rank influential nodes and compute shortest paths. Algorithms run on a graph projection with resource limits you set.

Algorithms

APOC utilities

Selected APOC functions and procedures are built in: traversal, data conversion, updates and import and export helpers.

APOC coverage

Streams and replicas

Opt-in change streams and subscriptions with label and property triggers, plus read-only replicas on the same instance or another server, including delayed replicas you can pause, resync or promote.

Change streams

Operations

Native TLS, statement and transaction time limits, listing and stopping running queries, PROFILE, and Prometheus metrics with a health endpoint.

Monitoring

Moving from Neo4j

Import offline Neo4j 4.x and 5.x dump files directly, from Explorer or with one procedure call, then point your existing drivers at Galactus DB.

Dump import
AI READY

Your graph, ready for AI agents.

The Galactus DB MCP server lets Claude and other AI assistants read your schema, run Cypher and search the docs, as a database user whose roles you control.

Connect an agent

How it runs.

The default Paged engine keeps the graph, its tables and every index on disk in checksummed pages, and holds the pages in use in memory inside one budget you set. A database opens straight from disk, and a graph can be larger than the server's RAM. Reads run together and never wait for writers.

Users and custom roles can grant access to individual databases, with separate system administration permissions. The protected system catalog stays in memory with durable backing.

Manage users and permissions

Each Galactus DB instance runs on one server, and replicas follow it on the same instance or, with Enterprise, on servers of their own. A delayed replica keeps a window of history in reserve, and any replica becomes a writable database with one PROMOTE statement.

Compatibility guide
YOUR APPLICATIONCypher and SQL · native Galactus DB driver
↓ Bolt 4.4 or 5 · TCP, TLS or WebSocket
GALACTUS DBQuery engine · tables · indexes · graph analytics
↓
DURABLE STORAGEChecksummed pages · commit log · one memory budget

Try it with your own data.

Community is free for any use and needs no account or licence key. One Compose file starts the database and Explorer.