> For the complete documentation index, see [llms.txt](https://docs.agenticflow.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.agenticflow.ai/reference/nodes/firecrawl_extract.md).

# Firecrawl Extract

**Action ID:** `firecrawl_extract`

## Description

Extract structured data from websites using natural language prompts or JSON schemas. This node uses Firecrawl's AI-powered extraction to intelligently parse webpage content and return specifically requested data fields.

## Provider

**Firecrawl**

## Connection

| Name                 | Description                                      | Required | Category  |
| -------------------- | ------------------------------------------------ | :------: | --------- |
| Firecrawl Connection | The Firecrawl connection to use for the extract. |     ✓    | firecrawl |

## Input Parameters

| Name                | Type    | Required | Default | Description                                                         |
| ------------------- | ------- | :------: | ------- | ------------------------------------------------------------------- |
| urls                | array   |     ✓    | -       | The URLs to extract from                                            |
| prompt              | string  |     -    | -       | The prompt to use for the extract                                   |
| schema              | object  |     -    | -       | The schema to use for the extract                                   |
| enable\_web\_search | boolean |     -    | false   | When true, extraction can follow links outside the specified domain |

<details>

<summary>View JSON Schema</summary>

```json
{
  "description": "Firecrawl extract node input.",
  "properties": {
    "urls": {
      "title": "URLs",
      "type": "array",
      "items": {"type": "string", "format": "uri"},
      "description": "The URLs to extract from."
    },
    "prompt": {
      "title": "Prompt",
      "type": "string",
      "description": "The prompt to use for the extract."
    },
    "schema": {
      "title": "Schema",
      "type": "object",
      "description": "The schema to use for the extract."
    },
    "enable_web_search": {
      "title": "Enable Web Search",
      "type": "boolean",
      "default": false,
      "description": "When true, extraction can follow links outside the specified domain."
    }
  },
  "required": [
    "urls"
  ],
  "title": "FirecrawlExtractInput",
  "type": "object"
}
```

</details>

## Output Parameters

| Name   | Type   | Description                           |
| ------ | ------ | ------------------------------------- |
| result | object | The output from the Firecrawl extract |

<details>

<summary>View JSON Schema</summary>

```json
{
  "description": "Firecrawl extract node output.",
  "properties": {
    "result": {
      "title": "Result",
      "type": "object",
      "description": "The output from the Firecrawl extract."
    }
  },
  "required": [
    "result"
  ],
  "title": "FirecrawlExtractOutput",
  "type": "object"
}
```

</details>

## How It Works

This node processes one or more URLs and extracts specific information using either a natural language prompt or a structured JSON schema. Firecrawl's AI understands the page content and extracts only the relevant data fields you request, returning results in the schema format you specify.

## Usage Examples

### Example 1: Extract Product Information with Prompt

**Input:**

```
urls: ["https://example.com/products/laptop-pro"]
prompt: "Extract the product name, price, specs, and customer rating"
enable_web_search: false
```

**Output:**

```
result: {
  "product_name": "Laptop Pro 15",
  "price": "$1299.99",
  "specs": {
    "cpu": "Intel i9",
    "ram": "32GB",
    "storage": "1TB SSD"
  },
  "customer_rating": 4.8
}
```

### Example 2: Extract Contact Info with Schema

**Input:**

```
urls: ["https://example.com/contact"]
schema: {
  "type": "object",
  "properties": {
    "company_name": {"type": "string"},
    "email": {"type": "string"},
    "phone": {"type": "string"},
    "address": {"type": "string"}
  }
}
enable_web_search: false
```

**Output:**

```
result: {
  "company_name": "Example Corp",
  "email": "contact@example.com",
  "phone": "+1-555-123-4567",
  "address": "123 Main St, City, State 12345"
}
```

### Example 3: Multi-URL Extraction with Web Search

**Input:**

```
urls: ["https://startup.example.com/about", "https://startup.example.com/team"]
prompt: "Extract founder names, founding year, and company mission"
enable_web_search: true
```

**Output:**

```
result: {
  "founders": ["Jane Smith", "John Johnson"],
  "founding_year": 2020,
  "mission": "To revolutionize digital transformation..."
}
```

## Common Use Cases

* **Data Extraction**: Extract structured data from unstructured web pages
* **Lead Enrichment**: Extract company information, contact details, and industry data
* **Product Information Gathering**: Collect product specs, pricing, and availability across websites
* **Research Data Collection**: Extract research papers, citations, and academic information
* **Real Estate Data**: Extract property details, pricing, and features from listings
* **Job Listing Analysis**: Extract job requirements, salary ranges, and qualifications
* **E-commerce Intelligence**: Extract competitor pricing, product features, and reviews

## Error Handling

| Error Type           | Cause                                                        | Solution                                                             |
| -------------------- | ------------------------------------------------------------ | -------------------------------------------------------------------- |
| Invalid URL          | URL format is incorrect or domain doesn't exist              | Verify all URLs are valid and properly formatted                     |
| Access Denied        | Website blocks automated scraping or requires authentication | Check if web search is enabled; consider authentication if available |
| No Data Extracted    | Prompt is too vague or schema doesn't match page content     | Refine the prompt or schema to match actual page structure           |
| Invalid Schema       | JSON schema is malformed or has syntax errors                | Validate schema against JSON schema standards                        |
| Multiple URLs Failed | Some or all URLs are inaccessible or time out                | Check URL availability and reduce page complexity                    |
| Web Search Error     | Web search feature encounters issues following links         | Disable web search and specify exact URLs needed                     |

## Notes

* **Multiple URLs**: You can extract from multiple URLs in a single request. Results will aggregate data from all pages.
* **Prompt vs Schema**: Use prompts for flexible, natural language extraction. Use schemas for consistent, structured output format.
* **Web Search**: Enable web search to allow the extractor to follow external links if information isn't on the specified pages.
* **Schema Format**: Provide JSON schema that matches standard JSON schema conventions for best results.
* **Rate Limits**: Respect Firecrawl's rate limits, especially when extracting from many URLs or using web search.
* **Accuracy**: More specific prompts and schemas yield more accurate extraction results.