How to Use ParadeDB on AWS in PostgreSQL

How do I install and use ParadeDB on AWS PostgreSQL instances?

ParadeDB on AWS adds fast vector search to Amazon-hosted PostgreSQL by installing the ParadeDB extension, letting you store embeddings and run similarity queries.

Welcome to the Galaxy, Guardian!
You'll be receiving a confirmation email

Follow us on twitter :)

Oops! Something went wrong while submitting the form.

Description

Example H2

Example H3

What is ParadeDB and why run it on AWS?

ParadeDB extends PostgreSQL with a native vector type and ANN indexes. Running it on Amazon RDS or Aurora gives you managed backups and scaling while keeping vectors close to transactional data.

How do I install ParadeDB on Amazon RDS?

Create a custom parameter group that adds paradedb to shared_preload_libraries, reboot, then run CREATE EXTENSION paradedb; from an rds_superuser role.

How do I create the extension and roles?

After reboot, connect with the master user and execute:

CREATE EXTENSION IF NOT EXISTS paradedb; GRANT USAGE ON SCHEMA paradedb TO app_user;

How do I store embeddings in my ecommerce schema?

Add a vector column to products to hold a 768-dimensional embedding:

ALTER TABLE products ADD COLUMN embedding vector(768);

How do I build an ANN index?

Load vector data first, then create IVFFlat or HNSW:

CREATE INDEX idx_products_embedding ON products USING ivfflat (embedding vector_cosine_ops) WITH (lists = 100);

How do I run a similarity query?

Find five products closest to a query vector:

SELECT id, name, price FROM products ORDER BY embedding <-> '[0.12,0.03,...]'::vector LIMIT 5;

Best practices for ParadeDB on AWS

Choose r6g or m6g instances for more RAM. Keep work_mem ≥ 64MB for vector sorts. Monitor paradedb_ivfflat_searches_total in CloudWatch.

Common mistakes to avoid

Lack of superuser role: RDS blocks CREATE EXTENSION for normal users. Use rds_superuser or ask AWS Support to enable the extension.

Index too early: Populate most embeddings before building IVFFlat; otherwise you’ll need REINDEX later.

Is ParadeDB production-ready on AWS?

Yes. ParadeDB is open-source and battle-tested. Enable automated snapshots, run nightly VACUUM ANALYZE, and keep ParadeDB version in sync with major Postgres upgrades.

Why How to Use ParadeDB on AWS in PostgreSQL is important

How to Use ParadeDB on AWS in PostgreSQL Example Usage


-- Recommend similar products to the one a customer is viewing
WITH target AS (
  SELECT embedding FROM products WHERE id = 101
)
SELECT p.id, p.name, p.price
FROM products p, target t
WHERE p.id <> 101
ORDER BY p.embedding <-> t.embedding
LIMIT 4;

How to Use ParadeDB on AWS in PostgreSQL Syntax


-- Enable ParadeDB extension (RDS/Aurora)
CREATE EXTENSION IF NOT EXISTS paradedb;

-- Add vector column to Products table
ALTER TABLE products ADD COLUMN embedding vector(768);

-- Populate embeddings (example value truncated)
UPDATE products
SET embedding = '[0.11,0.02,...]'::vector
WHERE id = 42;

-- Create IVFFlat index after data load
CREATE INDEX idx_products_embedding
  ON products USING ivfflat (embedding vector_cosine_ops)
  WITH (lists = 100);

-- Similarity search using cosine distance
SELECT id, name, price
FROM products
ORDER BY embedding <-> '[0.12,0.03,...]'::vector
LIMIT 5;

Common Mistakes

Trying to run CREATE EXTENSION without rds_superuser privileges. RDS blocks superuser commands. Fix: connect as the master user or grant the rds_superuser role before installation.
Building an IVFFlat index before vectors are loaded. The index needs representative data for good clustering. Fix: insert at least 10,000 vectors, then create the index.