Reads up to 50 candidates on a position in one request, each with any of their questionnaires, documents, scorecards, conversation, custom fields and activity stream. This is a read: nothing is changed. Use it in place of one GET /candidate/{candidateId} call plus one call per extra for each candidate.

Send exactly one of:

• candidate_ids: 1 to 50 candidate ids on this position. Ids that aren't on the position are listed in not_found.
• stage_id: every candidate currently in that pipeline stage, most recently updated first, one page at a time with page (0-based) and page_size (up to 50). Use the stage id values from GET /company/{companyId}/pipeline/{pipelineId}. Read the next page until a page returns fewer than page_size candidates.

Each candidate has the same fields as GET /candidate/{candidateId}, and each extra has the same fields as its own endpoint, with these limits per candidate: 10 questionnaires, 20 documents, the first page of 50 conversation messages and the first page of 50 stream events.

The same permissions apply as when each candidate is read on its own. A candidate the caller can't see (in a restricted stage, or private) is listed in denied with the reason instead of failing the request. An extra the caller isn't allowed to read comes back as an empty array, and without permission to see the team's scorecards only the caller's own are returned.