Skip to content

Contrast vs PostgreSQL

PostgreSQL runs your app. Contrast streams its changes into a warehouse, so dashboards and agents scanning a year of orders stop sharing a server with your app or waiting on a replica you have to size.

Updated November 2026

At a glance

$0
for idle compute
$10
per TB scanned
100%
tables in Apache Iceberg

Figures from the Contrast pricing page and docs.

How you're billed

List prices as each vendor publishes them, checked November 2026.

Contrast

Billed on
Data scanned
Price
$10 per TB

50 GB a month free. No monthly fee.

PostgreSQL

Billed on
The servers it runs on
Price
Free software

You pay for the servers, reporting replicas included.

Running dashboards on the database your app depends on?

Analytics reads far more rows than your app does. Here's where each one puts that work.

Contrast

Embrasure Flow writes columnar Parquet files and commits them to Apache Iceberg, which other engines can read too.

PostgreSQL

Whole rows on 8 kB pages, in the data directory of a server you run.

Moving analytics off your primary

Both read your primary through logical replication. PostgreSQL copies into a second server you keep in schema sync by hand; Contrast writes Iceberg tables you query by data scanned.

Contrast
contrast source connect postgres \
  --credential-env DATABASE_URL

contrast ingest plan <connection-id> \
  --source-kind postgres
contrast ingest start <connection-id> \
  --source-kind postgres
PostgreSQL
-- on the primary
CREATE PUBLICATION analytics
  FOR TABLE orders, customers;

-- on a second server, after pg_dump --schema-only
CREATE SUBSCRIPTION analytics
  CONNECTION 'host=primary dbname=app user=repl'
  PUBLICATION analytics;

Contrast and PostgreSQL, feature by feature

ContrastPostgreSQL

Storage

Columnar storageContrastYes. Parquet files in Apache Iceberg.PostgreSQLNo. Whole rows on 8 kB pages.
Tables other engines read directlyContrastYes. Any Iceberg-compatible engine can read them.PostgreSQLPartly. Only through a PostgreSQL connection.

Ingestion

Inserts, updates and deletes streamedContrastYes. Each change written as it commits.PostgreSQLYes. Logical replication sends every change.
New columns carried to the copyContrastYes. Added columns follow the source.PostgreSQLNo. Schema changes aren't replicated.
Open source change captureContrastYes. Embrasure Flow is Apache-2.0.PostgreSQLYes. Released under the PostgreSQL License.

Compute and pricing

Analytics off your app databaseContrastYes. Queries run on Contrast compute.PostgreSQLPartly. On a replica you run.
No server to size or runContrastYes. Serverless needs no instance.PostgreSQLNo. You run and size every server.
Billed by data scannedContrastYes. $10 per TB after 50 GB free.PostgreSQLNo. Free software on servers you pay for.
Runaway query limits on by defaultContrastYes. 500 GB and $100 a day to start.PostgreSQLNo. statement_timeout is off by default.

Agents and BI

PostgreSQL wire protocolContrastYes. pg.embrasure.ai, in alpha.PostgreSQLYes. Its native protocol.
Read-only agent SQL by designContrastYes. The MCP tool and endpoint only read.PostgreSQLPartly. Depends on the grants you set.

Running alongside PostgreSQL

  1. Set wal_level to logical (rds.logical_replication on Amazon RDS) and grant the ingestion role REPLICATION and SELECT.
  2. Add a PostgreSQL source with the database's direct endpoint, pick tables and run preflight. It names any missing setting, permission or primary key.
  3. Start the sync. Contrast copies each table, then keeps it current while your app keeps writing to Postgres.
  4. Connect agents with contrast setup, and point BI tools at pg.embrasure.ai on port 5432 with a personal access token.

More comparisons

PostgreSQL details are from its public documentation, checked November 2026: Database page layout, Database file layout, Parallel query, Logical replication, Logical replication restrictions, Client connection defaults, Frontend/backend protocol, PostgreSQL License. Product names are trademarks of their owners.