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

# PollJob

> Poll async job status and retrieve results.

Poll the status of an async job and retrieve results when complete. The agent decides when and how often to poll -- the tool never blocks.

## When to Use

* Check the status of a running comparison or parse job
* Retrieve the list of generated overlays after a comparison completes
* Report progress to the user while a job is processing
* Short-poll with `wait_seconds` to reduce round-trips

## Parameters

| Parameter      | Type   | Required | Description                                                                                               |
| -------------- | ------ | -------- | --------------------------------------------------------------------------------------------------------- |
| `job_id`       | string | yes      | Job ID returned by [`Compare`](/mcp/tools/compare) or [`Parse`](/mcp/tools/parse)                         |
| `wait_seconds` | number | no       | Short-poll: wait up to this many seconds for a status change (max 15). Reduces round-trips for fast jobs. |

## Response

### In Progress

```json theme={null}
{
  "id": "job_cmp_01",
  "status": "started",
  "progress": {
    "completed": 3,
    "total": 8
  }
}
```

| Field                | Type   | Description                               |
| -------------------- | ------ | ----------------------------------------- |
| `id`                 | string | Job ID                                    |
| `status`             | string | `"queued"`, `"started"`, or `"completed"` |
| `progress.completed` | number | Number of sheet pairs processed so far    |
| `progress.total`     | number | Total sheet pairs to process              |

### Completed

```json theme={null}
{
  "id": "job_cmp_01",
  "status": "completed",
  "progress": {
    "completed": 8,
    "total": 8
  },
  "result": {
    "overlays": [
      {
        "id": "ov_01",
        "block_a": {
          "type": "Plan",
          "sheet_number": "A-101",
          "drawing_name": "Arch Set Rev A"
        },
        "block_b": {
          "type": "Plan",
          "sheet_number": "A-101",
          "drawing_name": "Arch Set Rev B"
        },
        "score": 0.82,
        "uri": "storage://overlays/ov_01/composite.png"
      }
    ]
  }
}
```

The `score` indicates how similar the two blocks are (0.0 = completely different, 1.0 = identical). Lower scores mean more changes. After completion, use [`Query`](/mcp/tools/query) with `entity: "block"` and `include: ["overlays", "changes"]` on either block to get detailed change and clash descriptions.

### Failed

```json theme={null}
{
  "id": "job_cmp_01",
  "status": "failed",
  "error": {
    "code": "job_failed",
    "message": "Comparison failed: no matching sheets found between drawings"
  }
}
```

## Examples

### Poll a running comparison

```json theme={null}
{
  "job_id": "job_cmp_01"
}
```

If the job is still running, the agent can report progress to the user: "3 of 8 sheet pairs processed. Checking again..."

### Retrieve completed results

```json theme={null}
{
  "job_id": "job_cmp_01"
}
```

When `status` is `"completed"`, the `result.overlays` array lists every matched block pair with its similarity score. The agent can sort by score to identify sheets with the most changes and drill into them first.

<Tip>
  A good polling pattern: wait 2-3 seconds after triggering `Compare`, then poll. For large drawing sets, the agent can tell the user "Processing 12 sheet pairs, this may take a minute" and poll at longer intervals.
</Tip>

## Related Tools

* [`Compare`](/mcp/tools/compare) - Trigger a comparison job
* [`Parse`](/mcp/tools/parse) - Trigger a parse job
* [`Query`](/mcp/tools/query) - Fetch detailed changes, clashes, or extracted features after the job completes
* [`ViewImage`](/mcp/tools/get-image) - View overlay images referenced in the results