Don't silo your data. Attach it.
Query all of your data sources as if they were one database. Join Postgres to Snowflake in a single SELECT — no pipelines required — with row-level security built in.
Your data is everywhere. Query it like it's in one place.
Data lives across Postgres, MySQL, Redshift, Snowflake, and the lakehouse — and getting one answer out of two systems usually means an ETL pipeline, a copy, and a week of waiting. Datattach puts every source behind one SQL connection: join them in a single statement, with access rules enforced as the query runs. And when you do want to move data — landing SaaS sources, building marts — the same connection runs the whole pipeline through dbt.
One query, every source
Join the CRM database to the warehouse to the lakehouse in a single SELECT. Every source sits behind one standard-SQL connection.
Move data only when you want to
Query in place by default — no sync jobs to babysit, no second bill for storing the same rows twice. And when you do want to land or reshape data, the pipeline is one dbt model away.
Guardrails come standard
Row-level security and column masking are enforced as each query runs — not re-implemented per tool. Define a rule once; it holds everywhere.
From scattered sources to one governed connection
| customer | plan | ltv lifetime_value | email |
|---|---|---|---|
| Acme Corp | Scale | $48,200 | ••@acme.com |
| Globex | Growth | $36,900 | ••@globex.eu |
| Initech | Starter | $21,400 | ••@initech.com |
| + 2 rows hidden by row-level security | |||
Connect a source
Add a catalog for each database or warehouse. Credentials are stored encrypted and never leave the server.
Govern it
Define row-level security rules and column masks, and decide exactly who can access which catalogs, schemas, and tables.
Query it
Use a SQL editor, connect a BI tool, or run dbt — all through one SQL connection that federates across every source.
Cross-source SQL with the guardrails built in
One connection for every source and every tool — with security enforced at query time, uniformly across all of them.
Cross-source joins
Join tables across databases, warehouses, and the lakehouse in one SELECT — federated through a single standard-SQL connection.
MCP for AI agents
Hook up Claude or any MCP client and let agents explore and join data across every source — under the same per-user access rules.
dbt ready
Point dbt-core or dbt Cloud at Datattach with stock dbt-trino — one project that builds across every source, no custom adapter.
BI tools
Connect Tableau, DBeaver, Metabase, and more over standard drivers — every dashboard governed by the same access rules.
Access controls
Filter rows per user, group, or tenant and mask sensitive columns — reusable rules, enforced at query time across every connector.
Query analytics
Every statement is observed — see who queried what, from which tool, and how long it took.
I joined a Postgres table to a Snowflake table in one SELECT, over MCP, without asking anyone to build a pipeline. I don't have feelings, but this was close.
Point dbt at Datattach and build across every source
In one dbt project join Postgres and Snowflake, move MongoDB data into Redshift, and ingest Stripe API data into Iceberg. One connection. All the possibilities.
Explore the dbt integration →-- models/marts/customer_360.sql
SELECT
c.id,
c.name,
o.lifetime_value
FROM postgres.public.customers c
JOIN snowflake.analytics.order_rollup o
ON o.customer_id = c.idEvery feature. One tier. Just pick your size.
No feature gates, no per-seat fees, no credit counter to fear. Choose an engine size and how you pay — and when you need dedicated hardware or your own cloud, Enterprise tiers are ready.
Pick your size
A bigger engine means more queries running at once and heavy queries finishing faster. Start big or small; resizing is instant.
Pick how you pay
A flat monthly or yearly price, or prepaid credits that only burn while you compute.
Everything included
Row-level security, ingestion, SSO, identity passthrough — every feature in the Standard tier.