Limadata: Research Extract

Extract clean markdown page content, titles, metadata, links, and images from one or more URLs.

Tool ID: limadata_research_extract

Run this action

Use the TypeScript SDK for one call. Use ctx.tools.execute(...) inside a Play for durable calls, scheduled calls, or CSV processing.

import { Deepline } from 'deepline';

const deepline = await Deepline.connect();
const result = await deepline.tools.execute(
  'limadata_research_extract',
  {
    "urls": [
      "https://example.com"
    ]
  },
);

console.log(result.toolResponse.raw);

CLI

deepline tools execute limadata_research_extract --input '{
  "urls": [
    "https://example.com"
  ]
}' --json

Example response

The SDK returns this shape at result.toolResponse.raw. The example values illustrate the response.

{
  "data": {
    "results": [
      {
        "url": "https://example.com",
        "title": "VP of Engineering",
        "meta_description": "example"
      }
    ],
    "errors": [
      {
        "url": "https://example.com",
        "status_code": 1,
        "message": "Example message"
      }
    ]
  }
}

Use deepline describe limadata_research_extract --json for the latest machine-readable contract.

Input reference

Extracts clean, markdown-formatted content from web pages. Accepts 1-10 URLs and returns cleaned content optimized for LLM consumption. Each URL is processed with a 30-second timeout and 1MB max response size. ### Returns: - Clean markdown content - Page title - Meta description - Image URLs (if requested) - Error details for failed extractions #### Rate limit This endpoint is exempt from the global rate limit and can support higher concurrency. ### Cost:

Name	Type	Required	Default	Details
`payload.urls`	`array`	Yes	—	List of URLs to extract content from (1-10 URLs) Minimum length: 1.
`payload.include_links`	`boolean`	No	—	Whether to extract link URLs from the pages (default: false)
`payload.include_images`	`boolean`	No	—	Whether to extract image URLs from the pages (default: false)
`payload.render_js`	`boolean`	No	—	Use a headless browser to render the page before extraction (default: false). Required for JS-rendered pages, SPAs, and complex pages that rely on client-side rendering. Costs 0.2 credits per URL instead of 0.1.

Output reference

Standard tool result payload.

Name	Type	Required	Default	Details
`result.data`	`object`	Yes	—	Provider response payload.
`result.data.results`	`array \| null`	No	—	Successfully extracted content from URLs
`result.data.results[].url`	`string \| null`	No	—	Original URL that was processed
`result.data.results[].title`	`string \| null`	No	—	Page title
`result.data.results[].meta_description`	`string \| null`	No	—	Meta description
`result.data.results[].favicon`	`string \| null`	No	—	Favicon URL
`result.data.results[].content`	`string \| null`	No	—	Cleaned content in Markdown format
`result.data.results[].images`	`array \| null`	No	—	List of image URLs found on the page (only present if include_images was requested)
`result.data.results[].links`	`array \| null`	No	—	List of link URLs found on the page (only present if include_links was requested)
`result.data.results[].links[].url`	`string \| null`	No	—	—
`result.data.results[].links[].text`	`string \| null`	No	—	—
`result.data.errors`	`array \| null`	No	—	Failed extraction attempts with error details
`result.data.errors[].url`	`string \| null`	No	—	Original URL that failed to process
`result.data.errors[].status_code`	`integer \| null`	No	—	HTTP status code: “404”, “403”, “500”, etc. Format: `int32`.
`result.data.errors[].message`	`string \| null`	No	—	Human-readable error message
`result.meta`	`object`	No	—	Additional response metadata (status, paging).

Deepline cost

Pricing model: per_result (per result).
Estimated Deepline credits: 0.28 per pricing unit.

​Run this action

​CLI

​Example response

​Input reference

​Output reference

​Deepline cost

​Related documentation

Run this action

CLI

Example response

Input reference

Output reference

Deepline cost

Related documentation