Cortex API
string[] · minItems: 1 · requiredstringstringstringstringintegerstring
Harvest web pages into the knowledge base as markdown (MDHarvest / crawl4ai)
Import web pages as markdown
POST
http://localhost:8000
/api/web-import
Harvest one or more URLs into the knowledge base as clean markdown. Returns a task_id immediately; poll GET /api/tasks/{task_id} for progress. Requires manage permission and web crawling to be enabled (returns 404 when disabled).
Import web pages as markdown › Request Body
urlsAbsolute http(s) URLs to crawl
collection_idCollection to add the harvested documents to
content_filtercrawl4ai filter: 'fit' (readability, default), 'raw', or 'bm25'
queryRelevance query (only used by the 'bm25' filter)
Import web pages as markdown › Responses
Web import job accepted
task_idBackground task id to poll for progress
accepted_urlsNumber of URLs accepted into the job
message