> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.resemble.ai/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.resemble.ai/_mcp/server.

# Get Investigation Run

`GET https://app.resemble.ai/api/v2/agents/{uuid}/runs/{run_id}`

Retrieve a persisted Detect Agent run, including the replay transcript and the exact configuration and memory snapshots used by the investigation.

## Path Parameters

| Parameter | Type     | Description                                                            |
| --------- | -------- | ---------------------------------------------------------------------- |
| `uuid`    | `string` | Stable agent identifier returned in the `uuid` field by `GET /agents`. |
| `run_id`  | `string` | Run UUID returned by the `run_started` event or a run-list response.   |

```bash
curl --request GET \
  'https://app.resemble.ai/api/v2/agents/verify_document/runs/6b208dee-f515-48d7-a895-cd595c46aa6d' \
  -H 'Authorization: Bearer YOUR_API_KEY'
```

## Response

```json
{
  "success": true,
  "item": {
    "uuid": "6b208dee-f515-48d7-a895-cd595c46aa6d",
    "status": "completed",
    "result": {
      "verdict": {
        "forced": false,
        "excerpt": "The clip contains synthetic visual artifacts and conflicts with the claimed source."
      },
      "recommended_action": "reject",
      "confidence": 0.98
    },
    "inputs": {
      "query": "Is this clip authentic?",
      "filename": "clip.mp4"
    },
    "has_media": true,
    "created_at": "2026-08-11T17:30:00Z",
    "updated_at": "2026-08-11T17:31:12Z",
    "transcript": [
      {
        "type": "detect",
        "config_name": "Verify a Document or Receipt",
        "model": "MODEL_ID",
        "detect": {
          "uuid": "DETECT_UUID",
          "label": "fake",
          "score": 0.97,
          "media_type": "video"
        }
      },
      {
        "type": "final_verdict",
        "intelligence": "{\"recommended_action\":\"reject\",\"confidence\":0.98}",
        "forced": false
      }
    ],
    "config_snapshot": {
      "name": "Verify a Document or Receipt",
      "system_prompt": "Assess the authenticity and provenance of the submitted media.",
      "tier": "investigation"
    },
    "memory_before": "# Detect Harness Memory\n\n## Learned patterns",
    "memory_after": "# Detect Harness Memory\n\n## Learned patterns\n- Verify the original upload date before trusting repost captions.",
    "error": null,
    "media_url": "/rails/active_storage/blobs/redirect/.../clip.mp4"
  }
}
```

`media_url` is `null` when no uploaded primary media was attached. Runs created from public URLs keep the URL in `inputs.url`.

> **Note**
>
> Internal cost events are removed from `transcript`, and internal cost totals are removed from `result`, even though the complete accounting record remains stored server-side.

If the run does not belong to the specified active agent, the API returns `404 Not Found`.