From web page
to usable data.
Extract JSON, Markdown, text, and links from a URL. Render JavaScript when your source needs it.
This domain is for use in illustrative examples.
- Single URLs or 20-URL batchesProcess one URL at a time or up to 20 in a single request.
- Static, JavaScript, or autoChoose static HTML, JavaScript rendering, or let the API decide.
- Per-URL settlementEach URL is settled individually, including failed URLs.
Turn a webpage into content your app can use.
The source help page is on the left. Choose an output format and watch ApiClay extract the useful content for the reader on the right. Your application can then store, index, or display the result.
From your input to a useful result.
- Choose a sample input
- Apply the feature’s workflow
- Inspect the result
This example runs entirely in your browser with illustrative data. Use the real feature’s guide and listing to check supported inputs, access rules, and pricing.
Explore the example, then use the setup guide to try the real feature.
Project setup and what to check
Input: A permitted page URL and the output your application needs. Markdown is useful for readable documents; links help you discover related pages.
Workflow: Extract the page, then inspect the main content and source URL. Choose JavaScript rendering when important content only appears after the page runs.
Result: Content your application can clean, chunk, index, or store. Your own ingestion pipeline decides how to deduplicate pages and attach source metadata.
Check that the expected content is present. Try a small batch before expanding to up to 20 URLs per request, and account for browser or proxy charges.
One URL. The output
your app needs.
Markdown for AI pipelines
Get clean, readable content that works well in LLM and RAG workflows.
JSON DOM for structured page representation
Receive a structured representation of the page’s content and elements.
Links for navigation audits
Extract internal and external links for crawling, monitoring, or analysis.
Example: Get page content as Markdown
curl --get 'https://api.apiclay.com/v1/api/apiclay/extract' \
-H 'Authorization: Bearer YOUR_APICLAY_API_KEY' \
-H 'X-ApiClay-Html: markdown' \
--data-urlencode 'url=https://example.com'# Example Domain
This domain is for use in illustrative examples.
[Learn more](https://www.iana.org/domains/example)# Example Domain
This domain is for use in illustrative examples.
[Learn more](https://www.iana.org/domains/example)Start small. Scale with a clear contract.
- 1
Choose a URL
Provide a single URL or up to 20 URLs in a batch request.
- 2
Select the output
Get JSON, Markdown, text, or links. Use static, JavaScript, or auto.
- 3
Consume the response
Integrate the structured result into your application.
JavaScript when needed
Render JavaScript to capture dynamic content, or use static HTML for faster, simpler pages.
Streaming batches
Submit up to 20 URLs and receive results as a streaming NDJSON response.
URL list (up to 20)
https://example.com https://httpbin.org/html https://www.iana.org/domains/example ...
NDJSON stream
{ "url": "https://example.com", … }
{ "url": "https://httpbin.org/html", … }
{ "url": "https://www.iana.org/…", … }
...Your application
Ingest
Analyze
Store
Build
A web extraction API for applications and AI pipelines
A web page can contain the information you need inside navigation, markup, or JavaScript-loaded content. ApiClay's extraction endpoint turns a permitted target URL into a response that a script, data pipeline, or AI application can consume. You choose the representation instead of maintaining a separate converter for every downstream task.
Use direct extraction for a one-off request or small batch. When a web workflow needs a reusable public contract, publish it as a Website-to-API listing. When it needs thousands of input rows and a visible spending ceiling, run it through Live Websets. These paths share the gateway but serve different operational needs.
Match the output to the consumer
Markdown is useful for a readable document or retrieval pipeline.
How it works
Text removes formatting, links expose destinations, and a semantic tree retains a compact outline. JSON DOM output preserves page structure for deterministic processing. The output format controls representation; application-specific entity extraction still needs your own mapping or downstream logic.
Render JavaScript only when necessary
Static mode fetches the upstream response without running a browser.
How it works
JavaScript mode loads the page in a browser, while auto mode can promote a static fetch when dynamic rendering is required. Browser rendering makes SPA content accessible, with an additional compute fee when a browser actually runs.
Focus on the content that matters
CSS selector scoping can restrict extraction to a main article, results list, or another relevant subtree.
How it works
Output controls can exclude noisy elements. Narrower responses reduce the amount of irrelevant markup your application has to process and help keep AI context focused on the requested source material.
Process small batches with clear results
POST extraction accepts up to 20 URLs and streams newline-delimited JSON in input order.
How it works
Each URL has its own outcome and settlement, so a failure does not turn the entire batch into an opaque result. Use a stream parser rather than waiting for one large JSON array.
Where this fits into your work.
RAG document ingestion
Fetch approved documentation pages as clean Markdown, then chunk and index them in your own retrieval system. Keep the source URL alongside every stored document.
Catalog research
Gather page content or hidden XHR responses from permitted product URLs, then map the fields needed for a comparison or catalog workflow.
Link and content audits
Extract links or visible text from a set of pages to inspect navigation, source material, or changes in published content.
Start with a focused test.
Use these steps to set up URL Extraction and review the result before making it part of a larger workflow.
Get an API key →- 1
Choose your target URL
Use a buyer API key and send a GET request to /v1/api/apiclay/extract with the URL to fetch. Start with one representative page.
- 2
Select the output and render mode
Request JSON, Markdown, semantic tree, text, links, or captured XHR JSON. Use a CSS selector to focus extraction and choose static, JavaScript, or automatic rendering.
- 3
Consume a response or batch stream
Parse the returned data in your application. For a POST batch of up to 20 URLs, consume the NDJSON stream and handle each URL's result independently.
Understand each workflow’s cost.
Match the feature’s limits and billing model to the work you want to run.
Explore API listings →Extraction is settled per successful URL. Browser rendering, verified CAPTCHA solving, and metered proxies can add usage charges depending on the target and request controls.
Failed URLs are refunded independently within a batch. A batch size limit is separate from a spending limit: use Websets when you want a job-level hard credit cap and managed progress for a larger URL list.
Open an API listing for its Reference, Playground, and Pricing. Check the selected route’s allowances and applicable infrastructure charges before increasing usage.
FAQs
Does URL extraction require a published API listing?
No. The platform extraction endpoint accepts a URL directly. Publish a listing when you want reusable routes, a defined schema, buyer-facing documentation, or marketplace monetization.
Can it extract JavaScript-rendered pages?
Yes. Browser rendering is available for dynamic pages. Select JavaScript rendering or use auto mode. Accessibility of the target still depends on its behavior, access controls, and your configured proxy or session.
What happens when one URL in a batch fails?
The stream reports each URL independently, and failed URLs are refunded independently. Your consumer should inspect every result and decide which failed URLs, if any, to retry.
Can I process more than 20 URLs?
Use multiple extraction requests or create a Live Webset for a larger seed list. Websets add column mapping, job progress, pause/resume, exports, and a hard spending cap.
Make the web part of your data pipeline.
Get an API key and start extracting structured data from any URL.