Configure your request
Open Universal API in the dashboard. Enter a public HTTP or HTTPS URL and choose Standard or Residential proxy routing, an optional country, and the extraction format.
| Format | Result |
|---|---|
| Original HTML | The page HTML. |
| JSON with page data | An extended JSON response, including the provider’s page data. |
| Markdown | Page content converted to Markdown. |
| AI extraction | Structured JSON based on the properties you describe in the additional field. |
Browser and advanced settings
Use headless browser enables remote browser loading. Return page source returns the original page HTML before JavaScript rendering. When the browser is off, page-source, JavaScript and selector settings are ignored.
Advanced settings accept cookies as semicolon-separated name=value pairs, JavaScript to run on the target page, a CSS selector to wait for, and a 5–60 second provider timeout. JavaScript is encoded automatically. It runs at the provider, never in your dashboard.
Credits and limits
The connected admin sees the provider’s actual credit balance. Estimates depend on proxy type and rendering: Standard uses 1, 2 or 10 credits; Residential uses 25, 50 or 125. Target-specific rules can affect cost. AI extraction adds a variable charge that is only known after execution.
The internal connection permits one request at a time, up to 6 starts per minute and 30 per UTC day. It does not buy a subscription or restart a plan. If the provider rejects a request or runs out of credits, the result shows an error. Failed requests are not automatically retried.
Run and inspect
Select Run request to open the result viewer. The result shows the provider’s HTTP status, elapsed time and credit charge when reported. Code displays the response as text. Preview shows a static HTML view, formatted JSON or a simple Markdown reading view.
HTML previews block scripts, forms, links and external resources. They can look different from the original page. Copy response and Download retain the raw response. Responses over 2 MB are explicitly marked as truncated.
Closing the result view does not stop the request. Stop waiting disconnects your browser from the wait, but the provider may still finish and charge the request.
Use the generated code
Choose cURL, Python, Node.js, PHP, Java or C#. The example updates immediately with your settings and calls the authenticated Crawlhub request endpoint. Set CRAWLHUB_ACCESS_TOKEN in your private runtime environment using Copy session token. This is a short-lived account token, not a permanent API key. Keep it private and obtain a fresh token after expiry.
The server keeps the shared provider key private. The separate Proxy tab describes direct ScrapingAnt proxy configuration, which needs your own provider API key and its certificate configuration. A Crawlhub session token cannot authenticate that direct proxy mode. Use the HTTP API examples for the selected extraction format.
Save without saving secrets
Save draft stores the request settings on your account but excludes cookies and JavaScript. Download settings follows the same rule. Those advanced inputs remain only in the current tab. Generated code does include your current inputs, so check it before sharing.
Response bodies are not saved as account history. The internal service retains request identifiers, a settings hash, status, elapsed time and credit counts for operational checks; records older than seven days are cleaned during later requests.