---
title: "Start here: your first 15 minutes in Rig — Rig guide"
description: "Find your way around the Rig workspace: what each part of the left rail is for, why your role hides some of it, where settings live, and what live versus validated means for a migrated report."
canonical: "https://rig.so/guides/start-here"
format: markdown
---

Search guides, references, playbooks…⌘K

Browse guides

# Start here: your first 15 minutes in Rig

A map of the app before you do anything with it. Fifteen minutes from now you will know what every group in the left rail is for, where your settings live, why you cannot see everything a colleague can see, and what it means when someone tells you a report is live.

👋You need an approved account to follow along. If you sign in and land on a page saying your signup is pending, an admin in your organisation still has to approve you. There is nothing to fix from your side.

## The left rail

Everything in Rig hangs off the rail down the left-hand side. At the top is the Rig logo, which takes you home. Below it are three groups, and they are grouped by what you are trying to do rather than by feature.

- **Build** is where you use the product: **Chat** (`/chat`), **Data Apps** (`/data-apps`), **VDMs** and **Missions**.
- **Refine** is where you improve the answers Rig gives: **Context Inspector** (`/context`), **Metrics** (`/metrics`) and **Skills** (`/skills`).
- **Connect** is where data comes from and how it is kept fresh: **Connections** (`/connections`), plus **Pipelines** and **Access** for admins, and **Context History** if you have rights to edit context.

At the bottom of the rail is a smaller cluster: **Inbox** for things waiting on you, **Help**, which lists the product tours and links out to these guides, **Settings**, and the workspace switcher. Most people belong to exactly one workspace and never touch the switcher.

## What your role hides

The rail only shows what your role allows. Pipelines is admin-only. Context History needs context-edit rights. So if a colleague describes a link you do not have, that is your role rather than a bug, and the fix is an access change from an admin, not a reload. The same rule applies inside pages: a control you cannot use is usually one your role does not grant.

## Settings

Settings sits at the bottom of the rail, on the `/profile` route. The page is headed **Settings** and its own groups run down the left in two blocks.

- **Account** holds **Preferences** and **Privacy & Data**. These are yours alone.
- **Workspace** holds **Integrations**, **Sharing** and **Developer**. Admins also get **LLM & Billing**, plus links out to usage analytics and the access gateway.

The one card worth finding today is **MCP Setup**, under **Integrations**. That is how you connect Claude, Cursor or Codex to Rig. You do not need it yet, but it saves a hunt later.

## Tours and the command palette

Under **Settings** → **Preferences** there is a **Product tours** section with one row per tour: **Around Rig**, **Asking questions**, **Data apps**, **Connections**, and for admins **Access & roles**. Each row has a **Guide** link and a **Start** button, and each tour replays from the beginning every time. The same list appears under **Help** in the rail. **Around Rig** is the fastest map of the app and only highlights things.

A tour skips any step whose rail item your role hides, so a tour with nothing left to show will not open at all. That is expected rather than broken.

Then press **⌘K** (Ctrl+K on Windows). This is the command palette, and it searches pages, chats, metrics and data apps at once. It is the quickest route back to anything you have seen once, and it is worth building the habit early.

## Live vs validated

If Rig has migrated an existing report for you, live and validated are two different claims, and mixing them up is the single most common source of early confusion.

- **Live** means the report is connected to your warehouse and can run against it. Opening a data app shows its most recent saved run; it does not execute the queries again. Use **Run Now** to query the data currently in your warehouse.
- **Validated** means somebody has compared those numbers against your old system, across several reporting periods, and agreed they match.

Check three things separately: when the source last synced into your warehouse, when the report last ran (shown in the run selector), and whether its numbers have been validated. A new run can still use old source data if the source has not synced. Check Connections or ask your admin about source freshness.

A report can be live on day one and not yet validated. That is normal. Until it is validated, treat the numbers as a working draft, keep running your old report alongside it, and raise every difference you find. Differences are how validation gets done, so reporting one is useful work rather than a complaint. Ask which of your reports are validated and which are still in parity checking, and do not assume the answer.

## Glossary

- **Workspace** — one company's Rig account. The switcher sits at the bottom of the left rail.
- **Dataset** — the warehouse, or the part of one, that your workspace asks questions against.
- **Data app** — a saved page that runs SQL and renders a result: a dashboard, a tracker, a report.
- **Workflow** — the definition behind a data app: what runs, in what order, and when.
- **Context layer** — the metric definitions, business terms, joins and rules Rig uses to turn your question into correct SQL.

## Where to go next

Read these in order. The first four take you from your first question to answers in your own AI tools; the rest are there when the matching job lands on your desk.

1. [What Rig is](https://rig.so/guides/getting-started) — where Rig sits alongside the tools you already use.
2. [Ask your data](https://rig.so/guides/chat-basics) — get a governed answer, and see the SQL behind it.
3. [Data apps](https://rig.so/guides/data-apps-basics) — open one, run it, and find out where a number came from.
4. [Connect your AI tools](https://rig.so/guides/connect-ai-tools) — point Claude, Cursor or ChatGPT at your workspace.
5. [Work with your data in Claude](https://rig.so/guides/ask-in-claude) — prompts worth stealing once MCP is connected.
6. [Connect a source](https://rig.so/guides/rig-ingest) — bring data in and keep it fresh on a schedule.
7. [Automate a workflow](https://rig.so/guides/automate-workflow) — turn a repeated question into something that runs itself.
8. [Teach Rig your definitions](https://rig.so/guides/tribal-knowledge) — metrics and usage rules, so answers stop drifting.
9. [People and roles](https://rig.so/guides/admin-people-and-roles) — add a teammate and give them the right access.

Was this guide helpful?

Yes No

Step 2 of 10

---

- [This page as HTML](https://rig.so/guides/start-here)
- [Site map for language models](https://rig.so/llms.txt)
- [API and agent documentation](https://rig.so/developers)
