> ## Documentation Index
> Fetch the complete documentation index at: https://schemabrain.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# describe_entity

> The entity's bound table, identity column, and full column list — one round-trip.

<Note>
  **Layer:** Semantic layer
</Note>

The entity's bound table, identity column, description, and full column
list. One round-trip gives the agent everything it needs to compose SQL
against the entity.

```json theme={null}
{
  "name": "customer",
  "description": "A registered user who can place orders.",
  "qualified_table": "public.users",
  "identity": "id",
  "origin": "manual",
  "columns": [
    {"name": "id", "data_type": "bigint", "nullable": false, "description": "...", "pii_sensitivity": "public"},
    {"name": "email", "data_type": "text", "nullable": false, "description": "...", "pii_sensitivity": "public"}
  ],
  "token_estimate": 184
}
```

`pii_sensitivity` reflects the per-column tag set at index time by the
PII classifier — one of `public`, `internal`, `confidential`, or `pii`.
The parallel `pii_categories` list (e.g. `["contact"]`) gives the
underlying category names that drove that sensitivity. Block-on-sensitivity
routing flows through `get_metric`, which propagates `pii_categories`
from the classifier; this field exposes the same signal for read-time
UX choices.

Pass `name` as a bare identifier — `customer`, not `public.customer`. The
schema qualifier belongs on the bound table, not the entity name. Errors
surface as `unknown_name` with a `list_entities` recovery hint.