---
title: Retrieve URL
description: Extract structured content from a single web page through Firecrawl
---

## Installation

Start in a ChatJS app with `chat.config.ts` and `components.json`.
Set `FIRECRAWL_API_KEY` in your app environment before starting the server.
The installed descriptor declares this credential requirement.

```bash
npx @chat-js/cli@latest add retrieve-url
```

The command installs the server tool, renderer, and registration descriptor under
`tools/chatjs/retrieve-url/`, along with required dependencies and shared renderer
helpers. It then regenerates the app's tool registrations.

## Usage

Ask your chat model to retrieve a specific public page and provide its complete URL. Use a model that supports tool calling.

The tool accepts a `url` string using HTTP or HTTPS. It scrapes the page and
uses extraction when title, description, or content is missing. Successful
results include title, content, URL, description, and optional language.

## Behavior and limitations

This item retrieves a supplied page, rather than searching the web. Missing
credentials, invalid URLs, and provider failures produce error results.
Returned URLs omit query strings and fragments.

## Customize

Edit the installed `tool.ts` and `renderer.tsx` to change behavior or presentation.
See [Installation](./install) for updates and sync, and
[Authoring Tools](./authoring) to publish your own item.