Quick start

← Start

These six steps match Get started on your Workspace’s Overview, which checks off each step as you complete it. You can start with an empty Workspace, or try Monomer on sample data first.

No warehouse yet? Start with sample data

1. Sign in to Monomer with Google and open your Workspace’s Overview.
2. In Get started, choose Explore sample data. Monomer sets up Acme Cloud (sample), a fictional SaaS company with shared metrics and four dashboards, and opens it when it is ready.
3. Under Start with a dashboard, open Board metrics. The other dashboards cover revenue movements, customer segments, and product engagement.
4. Connect an agent, then paste the example prompt from Get started: Build a dashboard of weekly active accounts by plan from the shared metrics.

The sample’s data is read-only; you and your agent can still build dashboards and edit metrics. Only Organization Admins can set up the sample, and it counts toward your Organization’s Workspaces. When you are ready for your own data, select Connect your warehouse in the sample’s banner and follow the steps below.

Before you begin

• Sign in to Monomer with Google and open your Workspace. All workspaces in the workspace switcher lists every Workspace by Organization with your role; see Workspaces and the switcher.
• Check your role. Connecting a data source needs the Owner role; choosing tables, semantic models, and metrics need Editor; asking questions needs Explorer. If you joined a team, ask a Workspace Owner for the role you need. Workspace roles explains each one.
• Have the connection settings for a warehouse you are authorized to query, and a dedicated read-only warehouse account. Without one, start with sample data.
• Pick a question and a date range that exists in your data. Revenue is the example here; use a measure that matters to your business.

1. Connect a data source

1. Open Sources → Connectors and select Create connector.
2. Choose the source type and select Next.
3. Enter a display name and fill in the settings and credentials from your administrator, then select Create. Enter credentials in this form rather than in an agent conversation.
4. Catalog opens with Choose tables to load. The first list of catalogs or schemas confirms the connection.

Add a warehouse connection lists the settings for each warehouse and what to do if the connection fails. If your team already has a connector, use it instead of creating a duplicate.

2. Choose tables to load

1. For BigQuery, Snowflake, Databricks, and Trino, first choose the catalogs to use and select Continue.
2. In Choose tables to load, check the schemas or tables your question needs.
3. Select Save and refresh and wait for the refresh to finish.
4. Open a table and check that the expected columns appear.

See Select and refresh Catalog.

3. Describe a table as a semantic model

A semantic model says what one row of a table means and which fields to group by and add up.

In the appWith an agent

1. In Sources → Catalog, open the table and select Create semantic model.
2. Describe the row grain, add a Primary key, a Time field, and a Measure such as revenue with aggregation Sum.
3. Read the Check panel, then select Save.

Describe a table walks through each field.

Once your agent is connected (step 6), you can ask it to draft the model and review the result in Semantics:

In Monomer workspace [workspace name], inspect the [schema.table] table in
Catalog. Tell me what one row represents and propose a semantic model with
its primary key, time field, and measures. Ask me about anything you can't
determine from the data before saving it.

4. Turn a measure into a metric

A metric gives a measure a shared business name that people, dashboards, and agents reuse.

In the appWith an agent

1. Open the semantic model. Next to a measure without a metric, choose Create metric.
2. Check the name, such as revenue, add a label and a description of what it includes, and save.

Define a reusable metric covers ratios and other calculation types.

Create a revenue metric from the [model] model's revenue measure. Explain
what it includes and excludes, then query it for [a small known period] so
I can compare it with a trusted total.

5. Ask a question in Explore

In the appWith an agent

1. Open Analyze → Explore and choose Metrics.
2. Choose your metric. Under by, choose a time grain such as week; under over, set the date range.
3. Select Run and read the result under the query. Switch between Table, Chart, and SQL.

Compare a small period with a total you trust. An empty result is not automatically zero. See Analyze a metric and Work with results.

Use Monomer workspace [workspace name]. Analyze [business question] for
[start date] through [end date], compared with [comparison period].
Inspect the existing metrics first and reuse matching definitions.
Explain the metric, date field, timezone, and filters you used, and show
the run reference. Do not modify warehouse data.

6. Connect an agent

1. Select your name at the bottom of the sidebar and choose Connect an agent. Your settings → Agents shows this environment’s MCP server address and short steps for each client.
2. Add the server to your client and sign in to Monomer when it asks. Follow the guide for Claude, Codex, or another client.
3. Ask the agent to list your Workspaces and the metrics in this one, without changing anything.

The step is complete once your agent has used its connection. Its work appears under your name with the app you connected, for example via Claude, and on Overview as the app’s badge on your avatar.

Next: build a dashboard with your agent

Agents create and edit dashboards; you view them and review later changes in the app. After checking an analysis, paste:

Build a dashboard in the same Monomer workspace using the analysis and
metric definitions we just checked. Include KPI cards with
period-over-period changes, a time trend, and useful breakdowns. Add
filters for relevant dimensions. Run the dashboard queries, check them
against the analysis, report failed queries explicitly, and give me the
dashboard link.

The link the agent returns opens the dashboard’s draft review screen; check that the totals match the analysis. When you ask for changes later, the agent edits the dashboard’s draft: open the dashboard, choose More actions then Review draft (also offered as Draft has changes in Dashboards and under Needs attention on Overview), check the preview, then choose Apply. See Create a dashboard with an agent and Review and apply dashboard changes.

If a step fails, follow Troubleshoot a blocked task or contact Monomer with the failed step and a non-sensitive request reference.