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

# Get task

> Fetch a single task record. Poll this to track pending tasks to completion.

## Path parameters

<ParamField path="taskId" type="string" required>
  The task identifier returned by [`POST /v1/tasks`](/api-reference/tasks/submit).
</ParamField>

## Response — `200 OK`

<ResponseField name="taskId" type="string">
  Task identifier.
</ResponseField>

<ResponseField name="sessionId" type="string">
  The session this task belongs to.
</ResponseField>

<ResponseField name="endUserId" type="string">
  The end-user the task ran for.
</ResponseField>

<ResponseField name="task" type="string">
  The prompt as originally submitted.
</ResponseField>

<ResponseField name="status" type="string">
  `"pending"` | `"running"` | `"completed"` | `"failed"` | `"cancelled"`.
</ResponseField>

<ResponseField name="result" type="string | object">
  The final answer. String for most tasks; JSON object for tasks that return structured data or when `schema` was provided.
</ResponseField>

<ResponseField name="structured_content" type="object">
  Only present when the request included `schema`. The object conforming to that schema.
</ResponseField>

<ResponseField name="html_dump" type="string">
  Only present when the request had `include_html_dump: true`. Raw HTML of the final page.
</ResponseField>

<ResponseField name="totalSteps" type="number">
  Number of agent steps executed.
</ResponseField>

<ResponseField name="durationMs" type="number">
  Wall-clock duration of the task in milliseconds.
</ResponseField>

<ResponseField name="visitedUrls" type="string[]">
  Domains the agent navigated to, deduplicated in order.
</ResponseField>

<ResponseField name="stepSummaries" type="StepSummary[]">
  Per-step trace: `{ n, type, msg, success }`.
</ResponseField>

<ResponseField name="error" type="string">
  Populated only when `status == "failed"`.
</ResponseField>

<ResponseField name="createdAt" type="string (ISO 8601)">
  When the task was submitted.
</ResponseField>

<ResponseField name="updatedAt" type="string (ISO 8601)">
  Last status change.
</ResponseField>

<Info>
  Additional internal fields may be present on the response for observability and may change without notice. Rely only on the fields documented above.
</Info>

## Example

<CodeGroup>
  ```bash curl theme={null}
  curl "$API_BASE/tasks/7f2a..." \
    -H "Authorization: Bearer $API_KEY"
  ```

  ```python Python theme={null}
  task = client.tasks.get("7f2a...")
  print(task.status, task.result)
  ```

  ```typescript TypeScript theme={null}
  const task = await client.tasks.get("7f2a...");
  console.log(task.status, task.result);
  ```
</CodeGroup>

## Polling pattern

Both SDKs' `tasks.run(...)` wrap submit + poll in one call. If you're polling yourself:

<CodeGroup>
  ```python Python theme={null}
  submission = client.tasks.submit(end_user_id="alice", task="...")
  while True:
      task = client.tasks.get(submission.task_id)
      if task.is_terminal:
          break
      time.sleep(2)
  ```

  ```typescript TypeScript theme={null}
  const submission = await client.tasks.submit({ endUserId: "alice", task: "..." });
  while (true) {
    const task = await client.tasks.get(submission.taskId);
    if (["completed", "failed", "cancelled"].includes(task.status)) break;
    await new Promise(r => setTimeout(r, 2000));
  }
  ```
</CodeGroup>

Recommended poll interval: **2 seconds**. Lower than that wastes requests; higher than that adds user-visible latency.

## Errors

| Code  | Meaning                          |
| ----- | -------------------------------- |
| `401` | Missing `Authorization` header   |
| `403` | Task belongs to another business |
| `404` | Unknown `taskId`                 |