Skip to main content
Bulkgrid exposes both synchronous and asynchronous behaviors, and customers should design around that distinction.

​
Search is synchronous

POST /api/v1/search returns immediately in the response cycle. Your application should still set sensible HTTP client timeouts, but you do not need run polling for search.

​
Extraction and crawl workflows are asynchronous

For extraction, crawl, and deep crawl workflows:
  • the create request returns a run object
  • completion time depends on workload and source behavior
  • your application should poll status rather than waiting on a single long request
  • keep create-run request timeouts short enough to fail fast on network issues
  • use polling for job completion instead of long blocking HTTP calls
  • enforce a maximum total wait time in your own system
  • move long-running workflows to background jobs when user experience depends on responsiveness