Connect Metabase to Layerz

Your warehouse already answers the question — revenue by month, spend by category. Layerz pulls that saved question, maps its columns once, and lands the result as actuals on a structured model you can reforecast and export to clean Excel. Metabase stays your BI. Layerz turns its answers into a model.

Get started, free

Why connect Metabase

The number is already right in your BI: the revenue question, the spend-by-category dashboard. But the model that plans on top of it is a spreadsheet where someone re-types that number every month.

Layerz pulls the saved question directly. Each result row becomes a staged transaction, your column mapping normalises it, and your mapping rules route it to the right line of the model. Re-sync and the actuals follow the warehouse.

What you get

A model on your BI truth. The actuals branch reads what the question returns, not a copy.
One mapping, replayed forever. Columns and routing rules are set once and drive every sync.
Scenarios that hold. Branch a plan without touching the warehouse or the SQL.
Audit-ready Excel. Export any time. Never paywalled.

Honest limits

Layerz reads question results through the Metabase API. It never writes to Metabase or the warehouse.
Rows need a date and an amount to project; a single-value KPI card will not map.
Layerz models the output, it does not replace Metabase for exploration and dashboards.

How to connect Metabase

Open Sources

In the Layerz web app, open your model and go to Sources.

Add the Metabase connector

Add a connector and pick Metabase.

Enter your instance details

Paste your instance URL, an API key and the ID of the saved question to pull.

Map your columns

Tell Layerz which columns hold the date, the amount and the category. The mapping is saved on the connection and drives every sync.

Land the actuals branch

Layerz classifies each row through your mapping rules and lands the figures as an actuals branch. Sync whenever the data moves.

Prefer to let Claude do it? With the MCP plugin, ask your agent to connect the question and author the column mapping for you.

FAQ

What does Layerz read from Metabase?

Only the saved question you point it at. Its result rows are staged, normalised through your column mapping and classified by your mapping rules.

Does it query my warehouse directly?

No. Layerz calls the Metabase API and reads the question results. Metabase keeps owning the SQL and the permissions.

What shape should the question return?

Rows with a date and an amount, ideally plus a category column to route on. A monthly revenue-by-line query is the typical case.

Can Claude set it up for me?

Yes. The Layerz MCP plugin lets Claude connect the question, author the field mapping and land the actuals from the chat.