search_qualified_talents_by_filters tool starts a qualified search from structured Kalent filters plus qualification criteria. Kalent returns a castingId immediately. Ask the assistant to call get_qualified_search_result every 10–20 seconds while nextAction is wait.
The start is a call with its own quota, target (target.greens, credit cap target.maxCredits). Within that cap Kalent analyses talents in successive waves on its own, until this call finds target.greens green talents, the cap is spent, or the talent pool is exhausted. Two credits are charged per judged talent, whatever the verdict, and never for a judgement started after the call’s green quota was met; Kalent reserves min(workspace balance, target.maxCredits) when the call opens and returns unused reserved credits when each wave ends. To spend more afterwards, use continue_qualified_search.
This tool does not accept a webhook: poll get_qualified_search_result instead. Calls, waves, credits, and edge cases are explained on the qualified search overview.
Call this tool when you want structured Kalent filters (filters, filterType, isRequired, isExcluded, isExactMatch, radius, history). For a natural-language qualified search, use search_qualified_talents_by_prompt.
Filters use the same model as search_talents_by_filters and POST /v1/search/talents. This tool does not accept relatedSearchTransactionIds or nbToFetch.
The HTTP counterpart is POST /v1/search/talents/qualified.
Input
Response
The tool returns JSON with:Errors
The tool response is marked as an error when:- There is no MCP session
- API credits are insufficient (
Insufficient API credits: {available} available, {required} required.) - No qualification criterion remains
- The search times out (
The search request timed out. Please try again. (ref: …))