Crawler / Page
Choose fetch or render for page crawl.
Page crawl starts when you already have a URL. Use fetch for fast HTML retrieval, render for JavaScript-loaded pages, and screenshot when the same URL also needs visual evidence.
Mode selection
Match the crawl path to the page behavior
Fetch and render share the same page crawl endpoint, but they are different operational choices. If the useful content is already in HTML, use fetch. If the page depends on JavaScript or browser behavior, use render.
If you do not know the URL yet, start with page search and crawl only the strongest results.
| Mode | Best for | Cost |
|---|---|---|
fetch | Docs, articles, reference pages, blogs, and HTML-first content. | 2 credits base |
render | JavaScript apps, client-side routes, hydrated pricing pages, and browser-only state. | 10 credits base |
screenshot | Visual records attached to crawl results, audits, and support review. | 20 or 50 credits |
Related page workflows
Use page crawl as the extraction step
Known URL workflows often branch from search results, then return Markdown for machines and optional screenshots for human review.
Fetch API Fast, low-cost HTML retrieval for pages that do not need browser execution. Render API Browser-based extraction for JavaScript-loaded and client-rendered pages. Page Search API Find candidate URLs before deciding which pages to fetch or render. Screenshot API Add visual proof next to extracted Markdown when review matters.