Skip to main content
Auxilian columns use seven Postgres types. The same rules apply to the web app and the MCP tools.

​
Summary

null stores NULL in any nullable column.

​
text

Any string. Numbers and booleans are converted to their text form, so 42 stores "42". Objects and arrays are rejected. Use jsonb for structured values.

​
integer

Whole numbers from -2,147,483,648 to 2,147,483,647. Accepts a JSON integer such as 42, or a string of digits such as "42" or "-7". Decimals such as 4.5 are rejected. Values outside the range fail with an Invalid value error. For larger whole numbers, use numeric.

​
numeric

Exact decimal numbers with no fixed precision, suited to money and measurements. Accepts a finite JSON number such as 1499.99 or a numeric string such as "1499.99". Results return numeric values as strings, such as "1499.99", so no precision is lost in JSON. Convert them when you do arithmetic.

​
boolean

true or false. The strings "true" and "false" are also accepted. Other values, including 1, 0, "yes", and "no", are rejected.

​
timestamptz

A point in time. Send an ISO 8601 string with a time zone:
Postgres also accepts other common formats, such as 2026-11-03 (midnight UTC). Results always return ISO 8601 in UTC with milliseconds, such as "2026-11-03T09:00:00.000Z". A timestamptz column can default to now(), the time of the insert.

​
uuid

A UUID string such as "0b7c1f7e-6f0e-4d3a-9a51-2f7c3d1b8e44". Use uuid columns for foreign keys, since every table’s id is a UUID. See Foreign keys. A uuid column can default to gen_random_uuid(), a new random UUID for each row.

​
jsonb

Any JSON value: an object, array, string, number, boolean, or null.
In the web app, type jsonb values as JSON text, such as {"tags": ["fiction"]}. Filters can’t reach inside jsonb values. Use like or ilike on the JSON text, or store values you filter on in their own columns.

​
Size limits

​
Defaults

Only two defaults exist, and each requires a matching type: Defaults can be set through MCP tools. The web app’s table forms don’t set defaults.

​
Changing a column’s type

Types can’t be changed in place. Add a column with the new type, copy the values, then drop the old column.