Store

Roo-VectorDB

A PostgreSQL-native vector database for AI agents

A PostgreSQL extension that combines relational database features with vector search — vectors up to 16,000 dimensions, IVF-Flat indexing, and SQL WHERE-clause filtering, on infrastructure you already run.

PostgreSQL 16+PythonC/C++Apache-2.0 code

Apache-2.0 code · free Evaluation License up to 50,000 vectors per table · runs entirely on your own hardware

16,000
Max dimensions
Float32 and Float16 vector types
4
Distance metrics
L2, L1, cosine, negative inner product
50–300%
Multi-core gain
vs. single-core, across 4 test datasets

What's under the hood

Built on PostgreSQL

Runs as a PostgreSQL 16+ extension, so vector search lives alongside your relational data instead of in a separate system to keep in sync.

Native SQL filtering

Combine vector search with ordinary SQL WHERE clauses, with query optimization across both — no bolted-on filter language to learn.

IVF-Flat indexing

Approximate nearest-neighbor search with configurable nlists and nprobes, tuned for the accuracy/speed tradeoff your workload needs.

SIMD + OpenBLAS acceleration

Vector operations use SIMD instructions and OpenBLAS integration, with multi-threaded, configurable query execution.

Real-time index sync

Index updates happen with zero-latency synchronization, so newly written vectors are searchable immediately.

Shared memory architecture

Efficient shared-memory data access keeps large vector sets fast to query without duplicating data per connection.

Evaluation vs. commercial license

The Evaluation License is free and supports up to 50,000 vectors per table. GPU acceleration and full-scale limits are commercial-license only — reach out and we'll set it up for your workload.

Request commercial/GPU access

Get started

# install the .deb package on Ubuntu/Debian, or use the provided Docker setup

CREATE EXTENSION roovectorcpu;

See the README for the full install guide, table schema with roovector / roohalfvec columns, and distance-operator reference.

Let's talk about your agents

Whether it's agent memory, a no-code build, or in-house model serving, tell us what you're working on and we'll point you to the right place to start.

Free Architecture Review

We'll look at your current agent setup and flag where memory, tooling, or serving is the bottleneck

Custom Deployment Guidance

Tailored recommendations for your use case, team size, and hardware

Enterprise Support Options

24/7 support, priority updates, and dedicated engineering resources

Built for teams running agents in production

From open-source Lint-AI users to enterprises deploying custom inference — RooAGI covers the whole stack.

Request Information