> ## Documentation Index
> Fetch the complete documentation index at: https://student-213fb9fc.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# 事件类型

> Webhook 事件参考

<div id="payload-structure">
  ## 负载结构
</div>

所有 webhook 事件均采用以下结构：

```json theme={null}
{
  "success": true,
  "type": "crawl.page",
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "data": [...],
  "metadata": {}
}
```

| 字段         | 类型      | 描述                            |
| ---------- | ------- | ----------------------------- |
| `success`  | boolean | 操作是否成功                        |
| `type`     | string  | 事件类型（例如 `crawl.page`）         |
| `id`       | string  | 任务 ID                         |
| `data`     | array   | 与事件相关的数据（见下方示例）               |
| `metadata` | object  | 来自你在 webhook 配置中的自定义元数据       |
| `error`    | string  | 错误信息（当 `success` 为 `false` 时） |

<div id="crawl-events">
  ## 爬取事件
</div>

<div id="crawlstarted">
  ### `crawl.started`
</div>

在爬取任务开始处理时发送。

```json theme={null}
{
  "success": true,
  "type": "crawl.started",
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "data": [],
  "metadata": {}
}
```

<div id="crawlpage">
  ### `crawl.page`
</div>

在爬取过程中，每抓取到一个页面就会发送此事件。`data` 数组包含页面内容和元数据。

```json theme={null}
{
  "success": true,
  "type": "crawl.page",
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "data": [
    {
      "markdown": "# 页面内容……",
      "metadata": {
        "title": "页面标题",
        "description": "页面说明",
        "url": "https://example.com/page",
        "statusCode": 200,
        "contentType": "text/html",
        "scrapeId": "550e8400-e29b-41d4-a716-446655440001",
        "sourceURL": "https://example.com/page",
        "proxyUsed": "basic",
        "cacheState": "命中",
        "cachedAt": "2025-09-03T21:11:25.636Z",
        "creditsUsed": 1
      }
    }
  ],
  "metadata": {}
}
```

<div id="crawlcompleted">
  ### `crawl.completed`
</div>

在爬取任务结束且所有页面都已处理时发送。

```json theme={null}
{
  "success": true,
  "type": "crawl.completed",
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "data": [],
  "metadata": {}
}
```

<div id="batch-scrape-events">
  ## 批量抓取事件
</div>

<div id="batch_scrapestarted">
  ### `batch_scrape.started`
</div>

在批量抓取任务开始处理时发送。

```json theme={null}
{
  "success": true,
  "type": "batch_scrape.started",
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "data": [],
  "metadata": {}
}
```

<div id="batch_scrapepage">
  ### `batch_scrape.page`
</div>

针对批处理中每个被抓取的 URL 发送。`data` 数组包含页面内容和元数据。

```json theme={null}
{
  "success": true,
  "type": "batch_scrape.page",
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "data": [
    {
      "markdown": "# Page content...",
      "metadata": {
        "title": "Page Title",
        "description": "页面描述",
        "url": "https://example.com",
        "statusCode": 200,
        "contentType": "text/html",
        "scrapeId": "550e8400-e29b-41d4-a716-446655440001",
        "sourceURL": "https://example.com",
        "proxyUsed": "basic",
        "cacheState": "miss",
        "cachedAt": "2025-09-03T23:30:53.434Z",
        "creditsUsed": 1
      }
    }
  ],
  "metadata": {}
}
```

<div id="batch_scrapecompleted">
  ### `batch_scrape.completed`
</div>

在批次中的所有 URL 均处理完成后发送。

```json theme={null}
{
  "success": true,
  "type": "batch_scrape.completed",
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "data": [],
  "metadata": {}
}
```

<div id="extract-events">
  ## 提取事件
</div>

<div id="extractstarted">
  ### `extract.started`
</div>

当提取任务开始处理时发送。

```json theme={null}
{
  "success": true,
  "type": "extract.started",
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "data": [],
  "metadata": {}
}
```

<div id="extractcompleted">
  ### `extract.completed`
</div>

在提取操作成功完成后发送。`data` 数组包含提取的数据和用量信息。

```json theme={null}
{
  "success": true,
  "type": "extract.completed",
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "data": [
    {
      "success": true,
      "data": { "siteName": "示例网站", "category": "科技" },
      "extractId": "550e8400-e29b-41d4-a716-446655440000",
      "llmUsage": 0.0020118,
      "totalUrlsScraped": 1,
      "sources": {
        "siteName": ["https://example.com"],
        "category": ["https://example.com"]
      }
    }
  ],
  "metadata": {}
}
```

<div id="extractfailed">
  ### `extract.failed`
</div>

当提取失败时发送。`error` 字段中包含失败原因。

```json theme={null}
{
  "success": false,
  "type": "extract.failed",
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "data": [],
  "error": "提取数据失败：超时已超出",
  "metadata": {}
}
```

<div id="agent-events">
  ## Agent 事件
</div>

<div id="agentstarted">
  ### `agent.started`
</div>

当 Agent 任务开始执行时发送。

```json theme={null}
{
  "success": true,
  "type": "agent.started",
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "data": [],
  "metadata": {}
}
```

<div id="agentaction">
  ### `agent.action`
</div>

在每次调用工具（scrape、search 等）后发送。

```json theme={null}
{
  "success": true,
  "type": "agent.action",
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "data": [
    {
      "creditsUsed": 5,
      "action": "mcp__tools__scrape",
      "input": {
        "url": "https://example.com"
      }
    }
  ],
  "metadata": {}
}
```

<Note>
  `action` 事件中的 `creditsUsed` 值是目前累计使用 credits 总量的**预估值**。最终准确的 credits 消耗量仅在
  `completed`、`failed` 或 `cancelled` 事件中可用。
</Note>

<div id="agentcompleted">
  ### `agent.completed`
</div>

当 agent 成功完成时会发送该事件。`data` 数组包含提取的数据以及消耗的总额度（credits）。

```json theme={null}
{
  "success": true,
  "type": "agent.completed",
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "data": [
    {
      "creditsUsed": 15,
      "data": {
        "company": "示例公司",
        "industry": "技术",
        "founded": 2020
      }
    }
  ],
  "metadata": {}
}
```

<div id="agentfailed">
  ### `agent.failed`
</div>

当 agent 遇到错误时会发送该事件。`error` 字段包含失败原因。

```json theme={null}
{
  "success": false,
  "type": "agent.failed",
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "data": [
    {
      "creditsUsed": 8
    }
  ],
  "error": "Max credits exceeded",
  "metadata": {}
}
```

<div id="agentcancelled">
  ### `agent.cancelled`
</div>

当用户取消代理作业时发送。

```json theme={null}
{
  "success": false,
  "type": "agent.cancelled",
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "data": [
    {
      "creditsUsed": 3
    }
  ],
  "metadata": {}
}
```

<div id="event-filtering">
  ## 事件过滤
</div>

默认情况下，你会接收到所有事件。若只想订阅特定事件，请在 webhook 配置中通过指定 `events` 数组：

```json theme={null}
{
  "url": "https://your-app.com/webhook",
  "events": ["completed", "failed"]
}
```

如果你只关心任务是否完成，而不需要逐页级更新时，这会很有用。