What is RAT?

RAT is a self-hostable data platform that makes data engineering accessible to everyone. Write SQL (or Python), run pipelines, and query your data — all from a web-based IDE.

Why RAT?

• Anyone can data — If you can write SQL, you can build data pipelines
• Self-hostable — One docker compose up and you’re running
• Batteries included — Editor, scheduler, query console, lineage, quality tests — all built-in
• Apache Iceberg — Your data lives in open table format, not locked in a vendor
• Git-like isolation — Each pipeline run gets its own branch — bad data never reaches production

How It Works

RAT uses a medallion architecture (Bronze → Silver → Gold) to organize data transformations:

You write SQL pipelines that transform data between layers. RAT handles execution, scheduling, versioning, and quality testing.

Quick Example

Here’s a complete pipeline that deduplicates raw orders into a clean Silver table:

-- @merge_strategy: incremental
-- @unique_key: order_id
-- @watermark_column: updated_at
 
SELECT
    order_id,
    customer_id,
    total_amount,
    status,
    updated_at
FROM {{ ref('bronze.raw_orders') }}
{% if is_incremental() %}
WHERE updated_at > '{{ watermark_value }}'
{% endif %}

Platform at a Glance

Get Started

Architecture Overview

RAT runs as 7 containers orchestrated by Docker Compose:

Read more in the Architecture Overview.