The Scenario
A new data engineer just joined the team. She needs to understand what's in the production database — every table, every field, what the fields mean, what type they are. The documentation wiki has a page that was last updated fourteen months ago. You've been tasked with producing a current data dictionary from Metabase's metadata and putting it in a Google Sheet so she can reference it while she gets up to speed.
The bad version:
- Navigate to each database in Metabase's admin panel, click through to each table, and manually note field names, types, and descriptions into a spreadsheet.
- Realize Metabase has six connected databases and 140 tables, and that you've just committed yourself to several hours of clicking.
- Finish half of it before getting pulled into another ticket, leaving the sheet in a half-documented state that's arguably worse than no documentation at all.
The documentation doesn't exist because no one has time to build it manually. That's been true for fourteen months.
The Easy Way: One Prompt in SheetXAI
SheetXAI is an AI agent that lives inside your Google Sheet. It can query Metabase's metadata API — listing databases, tables, fields, types, and descriptions — and write everything into the sheet in one pass, organized however you specify.
List all databases in Metabase, then for each database list all tables and their fields with column name, data type, and description, writing everything into this sheet with one tab per database
What You Get
- One tab created per connected database, named after the database.
- Each tab has columns: Table Name, Field Name, Data Type, Semantic Type, Description.
- Every field Metabase knows about lands in the sheet — including fields with no description, which land with a blank in the Description column so they're easy to find and fill in.
- The structure is consistent across all tabs, so the new engineer can use a single filter to search across the whole dictionary.
What If the Data Is Not Quite Ready
You only need one specific table's schema
Fetch the complete schema for Metabase table ID 47 including all column names, types, and semantic types, and write them into columns A through C of this sheet starting at row 1 with headers
Useful for a focused onboarding session on one table, or for validating a schema before a migration.
The descriptions are mostly blank and you want to flag them
List all fields for Metabase database ID 2 and write field name, data type, and description into this sheet — in column D, write NEEDS DESCRIPTION for any field where the description is empty
The new engineer gets an actionable list of what to document, not just a wall of blank cells.
You need a count of how many fields each table has
Fetch the schema for all tables in Metabase database ID 2 and write one row per table into this sheet with columns: table name, number of fields, and a list of the first five field names
A useful summary view for a team getting oriented in a new codebase — quickly see which tables are wide and which are narrow.
Full dictionary plus data quality audit in one shot
List all databases and tables in Metabase, write every field's name, type, semantic type, and description into a tab named for each database, then in a Summary tab write one row per database showing database name, table count, field count, and number of fields missing a description
One prompt produces the working data dictionary and the gap analysis. The engineering lead sees both in the same sheet.
The pattern across all of these: ask for the structure you need, not just the raw metadata. SheetXAI handles the traversal and the layout.
Try It
Get the 7-day free trial of SheetXAI and open the Google Sheet you're using for your team's data documentation, then ask it to pull the full Metabase schema into tabs. Also useful: auditing your Metabase workspace inventory, and the hub overview on connecting Metabase to Google Sheets.