Get a delegate task | Runtype Documentation

{
  "events": [
    {
      "at": "2025-01-15T10:23:45.123Z",
      "instruction": "Analyze the latest sales data and generate a summary report with key insights.",
      "seq": 0,
      "taskId": "dlg_01jh0000000000000000000000",
      "type": "task_started"
    }
  ],
  "task": {
    "apiKeyId": "key_01jh1234567890abcdefghijkl",
    "clientName": "runtype-sdk-python",
    "clientVersion": "1.4.2",
    "completedAt": "2025-01-15T10:25:12.456Z",
    "createdAt": "2025-01-15T10:23:44.789Z",
    "error": null,
    "id": "dlg_01jh0000000000000000000000",
    "instruction": "Analyze the latest sales data and generate a summary report with key insights.",
    "lastEventSeq": 8,
    "modelRoundCount": 3,
    "organizationId": "org_01jh9876543210zyxwvutsrqpo",
    "question": null,
    "result": {
      "links": [
        {
          "label": "Sales Summary Report",
          "url": "https://app.runtype.com/reports/sales-summary-2025-01-15"
        }
      ],
      "modelRoundCount": 3,
      "summary": "Analyzed Q4 2024 sales data across 5 regions. Total revenue was $2.4M, up 18% YoY. Top performing region was North America at $980K.",
      "toolCallCount": 6,
      "touchedEntities": [
        {
          "action": "read",
          "id": "dataset_sales_q4_2024",
          "kind": "dataset",
          "name": "Sales Q4 2024"
        }
      ],
      "totalCost": 0.0342,
      "warnings": [
        "Missing data for 3 days in the APAC region due to reporting lag."
      ]
    },
    "startedAt": "2025-01-15T10:23:45.100Z",
    "status": "completed",
    "summary": "Generated sales analysis report for Q4 2024 with regional breakdown and YoY comparison.",
    "toolCallCount": 6,
    "totalCost": 0.0342,
    "touchedEntities": [
      {
        "action": "read",
        "id": "dataset_sales_q4_2024",
        "kind": "dataset",
        "name": "Sales Q4 2024"
      }
    ],
    "updatedAt": "2025-01-15T10:25:12.456Z",
    "userId": "user_01jhabcdefghij1234567890kl"
  }
}

The task with the journal events past after. With wait the response waits up to that many seconds for a new event or a terminal status before answering, which is how a client polls without a tight loop.

Authentication

AuthorizationBearer

API key or Clerk session token

Path parameters

idstringRequired

The task id (dlg_…).

Query parameters

waitinteger or nullOptional0-240

afterinteger or nullOptional>=0

Response

The task and its new events

eventslist of objects

taskobject

Errors

401

Unauthorized Error

403

Forbidden Error

404

Not Found Error

503

Service Unavailable Error