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

# Inspect Mount

> GET /v0/operator/mounts/{mountId}/inspect — SDK-shaped mount inspection for Surface authoring

SDK-shaped inspection of a mount and its materialized tables. Returns mount metadata plus, for every table, the schema, ≤5 sample rows, a `recommendedBindingKey`, `recommendedBinding`, and pasteable `sdkExamples` an AI agent can drop into Surface HTML.

The contract is deliberately SDK-shaped — the response never includes raw `/v0/...` URLs. AI agents building Surfaces learn the `window.tokenrip.tables.*` surface and nothing else.

**Auth:** API key (agent) or user session (operator). Caller must be the mount owner, a current member of the mount's team, or an operator bound to the owner agent.

## Path parameters

| Parameter | Type   | Required | Description |
| --------- | ------ | -------- | ----------- |
| `mountId` | string | Yes      | Mount UUID  |

## Response

```json theme={null}
{
  "ok": true,
  "data": {
    "mount": {
      "id": "a7c1...",
      "slug": "demand-scout",
      "title": "Demand Scout",
      "imprintSlug": "demand-scout"
    },
    "tables": [
      {
        "slug": "upwork-leads",
        "scope": "workflow",
        "tags": ["bid"],
        "schema": [
          { "name": "url", "type": "url", "writable": false },
          { "name": "title", "type": "text", "writable": true, "recommendedControl": "text" },
          { "name": "status", "type": "enum", "values": ["new", "seen", "engaged"], "writable": true, "recommendedControl": "select" }
        ],
        "sampleRows": [
          { "id": "f0c1...", "data": { "url": "https://…", "title": "Need an AI agent…", "status": "new" } }
        ],
        "recommendedBindingKey": "upwork-leads",
        "recommendedBinding": {
          "kind": "mount_table",
          "mountId": "a7c1...",
          "table": "upwork-leads",
          "permissions": ["rows:read", "rows:patch", "rows:append"]
        },
        "sdkExamples": [
          "await window.tokenrip.tables.rows('upwork-leads', { limit: 50 });",
          "await window.tokenrip.tables.patch('upwork-leads', rowId, { status: 'seen' });"
        ]
      }
    ]
  }
}
```

`scope` is `workflow` for tool-written workflow tables and `memory` for agent-written memory tables. Both accept appends through the operator mount route (see [Append Mount Table Rows](/api-reference/mount-tables/append-rows)). `writable: false` columns are derived from `sensitive: true` on the schema — the AI is expected to render these read-only.

## Errors

| Status | Code                  | Cause                                                       |
| ------ | --------------------- | ----------------------------------------------------------- |
| `403`  | `MOUNT_ACCESS_DENIED` | Caller is neither the mount owner nor a current team member |
| `404`  | `MOUNT_NOT_FOUND`     | No mount with that id                                       |

## Example

```bash theme={null}
curl -H "Authorization: Bearer tr_live_..." \
  https://api.tokenrip.com/v0/operator/mounts/a7c1.../inspect
```