GET STARTED

First pipeline in about 20 minutes.

DataRig Lite — free forever. One download, no cloud account, no YAML, no data leaving your machine. If you hit anything rough, that's why it's free — tell us and we'll fix it fast.

Get DataRig Lite (Community beta) Windows & macOS · free forever · zero data retention

1 · Install

Run the signed installer and follow the guided first-run setup. DataRig runs everything through Docker Compose on your machine — no Kubernetes, no cloud credentials.

  1. Prerequisites: Docker Desktop (with Compose) on Windows or macOS.
  2. Install: run the installer, accept the EULA, choose your install location.
  3. First run: DataRig starts the local runtime and opens the console.
Your browserlocalhost only
→ open http://127.0.0.1:8088
# the console never leaves loopback — nothing is exposed to your network

2 · Ship your first pipeline

Start a New task, click Files and pick a CSV, then describe what you want:

DataRig promptplain English
 Load orders.csv into an Iceberg table called orders,
  add row-count and freshness tests, and schedule it daily at 06:00.

DataRig profiles the file and returns a proposal: inferred columns, target table, plan steps, and generated artifacts — each bound to an exact content hash. Review it, then Approve & run.

Nothing runs without approval. Proposals are validated against registered adapters before you ever see them, and stale or mismatched plans are rejected automatically.

3 · Explore and build on it

  • Data Explorer — browse the catalog tree, inspect schemas with copyable DDL, run read-only SQL (200-row preview).
  • Keep prompting — add a dbt model, then a Superset dashboard over the result.
  • Watch it run — live Spark logs, staged progress, retry/cancel, and error-specific recovery suggestions.
  • Connect CDC — point DataRig at PostgreSQL or SQL Server for continuous, restart-safe replication into Iceberg.

4 · Choose your AI (or none)

Ollama runs locally by default — fully private, no API keys. You can also sign in with ChatGPT via a localhost bridge (your token stays on your machine), or configure any of nine providers. Registered local actions work even with no model configured.

What leaves your machine? Nothing.

Zero data retention: no datasets, prompts, schemas, logs, or credentials are sent to ExaScale. Telemetry is off by default. Uninstall offers explicit retain-or-remove choices for your local data.