---
title: "Datarelix Quick Start"
description: "Ask your database in English. Sign in, connect a database, run schema discovery, and get your first verified answer."
canonical: https://docs.datarelix.ai/quickstart/
---

# Datarelix Quick Start

import { Steps, CardGrid, LinkCard } from '@astrojs/starlight/components';

Sign in, connect a database, let discovery run, and ask your first question.

## Before you start

You'll need:

- A **Datarelix account** — create one at [app.datarelix.ai](https://app.datarelix.ai) with your email address and a password (at least 8 characters). No credit card.
- A **database** Datarelix can reach, in one of the supported engines: Postgres, MySQL, Azure SQL, Databricks, Kusto, BigQuery, Athena, Elasticsearch, or Snowflake — or data files in S3/Azure/GCS via the Files & Object Storage (DuckDB) connection.

Datarelix runs the execution layer, the per-engine query services, and the Python sandbox for you — there's nothing to install or operate.

## Your first query

<Steps>

1. **Create your account**

   Open [app.datarelix.ai](https://app.datarelix.ai). The sign-in page opens on **Create your account**: enter your name, email, and a password of at least 8 characters, then confirm it. You're signed in as soon as the account is created — a confirmation email follows, and you don't have to wait for it to start querying.

   Already have an account? Switch to **Sign in**, or use **Forgot password?** to get a reset link.

2. **Add a connection**

   Click **Add connection** and pick your database engine. Each engine has its own fields and auth modes — the [Connections guides](/guides/connections/) walk through every one, including how to create a read-only role and where to find each credential. Datarelix only ever reads inside the **allowed scope** you set (schema / database / dataset / index pattern).

3. **Let discovery run**

   After you connect, Datarelix introspects the schema and optionally enriches each table with a short description. Better descriptions mean sharper answers — see [Writing queries](/guides/queries/). How many tables one analysis covers depends on your plan; see [Billing & plans](/guides/billing/).

4. **Ask a question**

   Pick the connection and ask in plain language:

   ```text title="ask"
   What tables are in this database?
   Show me the top 5 customers by lifetime value
   What was the total revenue per month?
   ```

</Steps>

Every answer comes back artifact-first: a table or chart, the exact SQL/KQL that ran, the tables used, and the runtime — all one click away.

## Using the API

Every UI action is also a REST endpoint. Programmatic access uses an API key, which is **issued on request rather than self-serve** — there is no key-generation screen in the app yet. [Contact us](https://datarelix.ai/contact/) if you want one for your account. With a key, a run looks like this:

```bash
curl -N -X POST https://app.datarelix.ai/api/v1/run/stream \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $DATARELIX_API_KEY" \
  -d '{ "connection_id": "your-connection-id", "question": "What are the top 10 customers?" }'
```

See the [REST API reference](/api/rest/) for the full endpoint list and [Authentication](/api/auth/) for how sign-in and API keys work.

## Next steps

<CardGrid>
  <LinkCard title="Connections" href="/guides/connections/" description="Connect any supported engine." />
  <LinkCard title="Writing queries" href="/guides/queries/" description="Get sharper answers with table mentions and limits." />
  <LinkCard title="Python analysis" href="/guides/python/" description="Charts and stats in the sandboxed runtime." />
  <LinkCard title="Dashboards" href="/guides/dashboards/" description="Pin results and build updatable dashboards." />
  <LinkCard title="Security" href="/guides/security/" description="What the model sees, what it can run, and how your data is isolated." />
  <LinkCard title="Billing & plans" href="/guides/billing/" description="Free-plan limits, monthly counters, and upgrades." />
</CardGrid>