- create website preview jobs before a knowledge space exists
- persist crawled page content in KnowledgeFS object storage
- import selected pages using preview job and page IDs
- validate preview ownership, expiry, and configuration fingerprint
- clean up preview content after cancellation, expiry, or consumption
- add namespace preview routes, capabilities, migrations, and contracts
- keep online document and drive flows unchanged
- include Firecrawl configuration and selected page URLs in space creation
- start a server-side crawl import after KnowledgeFS space provisioning
- require the second crawl to match the complete user selection
- keep the knowledge space valid when website content import fails
- reuse existing background task status and failure UI
- add KnowledgeFS crawl-import API and regression tests