pg-vector-age

PostgreSQL 18 with pgvector, Apache AGE, Citus, and full-text search extensions — pre-configured, multi-arch, production-hardened.

What's inside

Component	Version	Purpose
PostgreSQL	18	Base
pgvector	0.8.2	Vector similarity search
Apache AGE	1.7.0	Graph queries (openCypher)
Citus	14	Distributed / columnar
pg_trgm	built-in	Trigram similarity + FTS indexes
unaccent	built-in	Accent-insensitive text search
btree_gin	built-in	Composite GIN indexes

Platform: linux/amd64 + linux/arm64 (native on Apple M-series, AWS Graviton)

Hardening

• scram-sha-256 — no trust or md5
• TLS — self-signed cert auto-generated on first start; mount your own at /var/lib/postgresql/tls/
• Audit logging — connections, disconnections, DDL (log_statement = 'ddl')
• Minimal attack surface — multi-stage build; no build tools in runtime image

Quick start

# docker-compose
cp docker-compose.example.yml docker-compose.yml
POSTGRES_PASSWORD=secret docker compose up -d

# connect
psql "host=localhost port=5432 dbname=app user=app sslmode=require"

Extensions

All extensions are created in the default database on first start:

-- vector similarity
SELECT '[1,2,3]'::vector <=> '[1,2,4]'::vector;

-- graph (openCypher via AGE)
SELECT * FROM ag_catalog.cypher('my_graph', $$
  CREATE (n:Person {name: 'Alice'}) RETURN n
$$) AS (n agtype);

-- full-text search
SELECT to_tsvector('english', 'The quick brown fox') @@
       to_tsquery('english', 'quick & fox');

-- trigram similarity
SELECT similarity('hello', 'helo');

-- Citus distributed table
SELECT create_distributed_table('my_table', 'tenant_id');

Custom init SQL

Mount additional init scripts:

volumes:
  - ./my-init.sql:/docker-entrypoint-initdb.d/02-my-init.sql

Bring your own TLS

volumes:
  - ./certs/server.crt:/var/lib/postgresql/tls/server.crt:ro
  - ./certs/server.key:/var/lib/postgresql/tls/server.key:ro

Build locally

docker build -t pg-vector-age .

License

Apache 2.0