Perplexity AI Answer Scraper with Sources

Perplexity AI HTML By URL

Fetches fully rendered Perplexity answer HTML and always opens a normalized Perplexity query URL.

The worker accepts either a Perplexity URL or plain search terms:

• {"url": "https://www.perplexity.ai/?q=apple"}
• {"search_terms": "apple"}

If url is provided, the worker extracts the query text from the URL and still opens https://www.perplexity.ai/?q=<search_terms>.

Input

At least one of url or search_terms must be provided.

Output

Behavior

1. Resolve input into search terms.
2. Open https://www.perplexity.ai/?q=<search_terms>.
3. Detect Cloudflare challenge pages with multiple signals.
4. Retry the full flow up to flow_retry_num times. When omitted, the default is 3.
5. Wait for the Perplexity answer to finish.
6. Scroll to the bottom, click Sources, and close follow-up popups when present.
7. Capture the answer stream response body as response_raw when Perplexity returns it.
8. Capture a PNG screenshot of the final visible page state as screenshot unless capture_screenshot is false.
9. Return gzip-compressed, base64 encoded HTML whenever available. If Cloudflare cannot be bypassed after retries, no HTML is returned.