Creating a database

Create a database in an app

1. Open Data Sheets from the dashboard sidebar and choose the app the data belongs to. Databases are team + app scoped, so every database lives under exactly one app.
2. Click New database. Give it a name and a slug — the slug is the stable handle and must be unique within the app (lowercase, hyphenated). Add an optional description.
3. Click Create. You land in the database with an empty sheet list, ready to add your first sheet.

How many databases you can create per app is plan-limited (max_databases_per_app; -1 means unlimited).

Add sheets

1. Click + Sheet (or the + on the sheet tab bar). A starter sheet with a couple of text columns is created.
2. Rename the sheet by double-clicking its tab. The slug regenerates from the name on first save and can be edited later — it's the identifier the SDK, workflow tasks, and agent tools use, so pick something stable.

A database can hold up to the plan's max_data_sheets_per_database sheets (default ceiling 20 when no plan override applies).

Define columns

• Click + Column to add a column, then set its display name and type. Types include text, number, boolean, date / datetime / time, large text, JSON, select / multi-select (with a fixed option list), array (free-form tags), the string variants url / email / phone, the number variants currency / percent / rating, and a system-managed auto increment id. See Managing sheets, columns, and rows for what each type does.
• Each column has a key (the field name code reads) derived from its name, and a display name you can change without breaking the key.
• Reorder columns by dragging the grip handle on a column header; the order is saved with the sheet. A sheet can hold up to 50 columns.

Add and import rows

• Click + Row to add an empty row and edit cells inline. Cell editors are type-aware; large text and JSON cells open a side editor. Edits auto-save.
• Import CSV to seed a sheet from a spreadsheet. Choose Append (add the CSV rows after the existing ones) or Replace (delete all current rows and insert the CSV — this clear is permanent). CSV headers match each column's key first, then its display name; unknown headers are skipped.
• Export CSV downloads the whole sheet with column keys as headers, so an export → edit → import round-trip is lossless.

Rows per sheet are plan-capped (max_data_sheet_rows). With the database created and sheets defined, link it from a playground (Using Data Sheets from a Playground), reference it from a workflow (Data Sheets in workflows), or attach it to an agent (Data Sheets in agents).