> For the complete documentation index, see [llms.txt](https://docs.uptiq.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.uptiq.ai/platform-resources/skill-library/ai-skills/web-crawler.md).

# Web Crawler

**Web Crawler** fetches a public URL and returns its content as both plain text and Markdown. It is how a workflow reads a page that isn't yours.

## How it works

* **Input resolution** — takes a `url` and an optional `instruction` as direct configuration.
* **Processing** — issues an HTTP GET to the URL, parses the HTML response to extract the main textual content, and formats it as both plain text and Markdown.
* **Execution** — blocking. Waits for the page to be fetched, extracted, and parsed.
* **Response** — success returns `extractedContent` and `markdown`; failure returns an error message and status code.

## Worked example: monitoring competitor pricing

You need to check a product's price on a competitor's public site regularly, rather than having someone look every Monday.

**The problem.** Browsing a page and copying a number out of it is inefficient, easy to get wrong, and nobody remembers to do it.

**The approach.** A Web Crawler skill fetches the page, and the `instruction` field narrows what comes back — *"Extract the pricing table for the Professional plan"* rather than the whole page including the navigation and the cookie banner.

**Configuration:**

```json
{
  "url": "https://www.example.com/news/article-123",
  "instruction": "Extract the main article text and headline."
}
```

**Output:**

```json
{
  "extractedContent": "Headline: Latest News. This is the main body of the article...",
  "markdown": "# Latest News\n\nThis is the main body of the article...",
  "statusCode": 200,
  "error": null
}
```

{% hint style="info" %}
**Two formats, different jobs.** `markdown` keeps the structure — headings, lists, emphasis — which usually helps a model understand what it's reading. `extractedContent` is flat text, better when you're matching or extracting rather than reasoning. Pick per use, and don't pass both to a model just because both are there.
{% endhint %}

## Configuration reference

| Field         | Type   | Required | Description                                                     |
| ------------- | ------ | -------- | --------------------------------------------------------------- |
| `url`         | string | ✅        | Absolute URL of the page to crawl.                              |
| `instruction` | string | —        | Guides extraction, e.g. *"Extract only the main article text"*. |
| `name`        | string | —        | Display name for this skill instance.                           |
| `description` | string | —        | Longer description shown in the designer.                       |

## Output

The output port (`crawledContent`) always conforms to:

| Field              | Type           | Always | Description                                  |
| ------------------ | -------------- | ------ | -------------------------------------------- |
| `extractedContent` | string         | ✅      | Plain text extracted from the page.          |
| `markdown`         | string         | ✅      | Markdown-formatted content from the page.    |
| `statusCode`       | number         | ✅      | HTTP-like status for the operation.          |
| `error`            | string \| null | —      | Error message on failure; `null` on success. |

## Errors

| Status | Meaning                                                             |
| ------ | ------------------------------------------------------------------- |
| `200`  | Successfully extracted data from the page.                          |
| `400`  | Bad request — malformed `url`, or a missing/invalid `instruction`.  |
| `500`  | Internal error — failed to parse or extract from the HTTP response. |

{% hint style="warning" %}
The skill reads **public** pages. It has no session, so anything behind a login, a paywall, or a bot check will not come back — and the failure may look like a successful fetch of a login page rather than an error. Check what you actually got, not just the status code.
{% endhint %}

## Testing in isolation

* **Path** — `/skill-runtime/workflows/nodes/WebCrawler/execute`
* **Method** — `POST`

***

To add this skill to an agent, see [Adding a Skill to the Agent](/agent-builder/build/adding-a-skill-to-the-agent.md). For search rather than fetching a known URL, see the MCP catalog under [MCP Tools](/agent-builder/build/mcp-tools.md).


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.uptiq.ai/platform-resources/skill-library/ai-skills/web-crawler.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.