Firecrawl crawl limit: 10,000 default vs Sume's 20 pages

Sume's crawl endpoint, POST /v1/firecrawl/crawl, accepts limit from 1 to 20 with a default of 10. Firecrawl's crawl guide says its default limit is 10,000 pages. A site section bigger than 20 pages needs several narrower crawls, or crawl_map to list URLs and crawl_scrape for the ones you want.

Firecrawl's numbers are from its crawl guide and Sume's from OpenAPI and MCP tool descriptions, read 2026-09-30.

How do the two crawl limits compare?

Both also differ in how you collect results.

How do I cover a site bigger than 20 pages?

Split it by section and give each crawl its own idempotency key, scoping each with include_paths (how they differ from Firecrawl's). Or call crawl_map (limit 1-100) for URLs and scrape the relevant ones.

How do I wait for a crawl?

The description says to call jobs_wait on the returned job_id, then crawl_get with the same id, and never to start a second crawl to poll. Reusing the key on a retry is what keeps a repeat from becoming a second crawl; see idempotency keys.

What does truncated mean?

crawl_get returns bounded pages, and truncated means the result was cut. The guidance is to narrow the scope if you need more content, which is the same fix as the 20-page limit: smaller prefixes, more crawls.