Skip to main content

RECRU connector

Back to the connector overview, authentication methods, and actions.
​
28 Oct 2026
breaking auth guide description-fixThe Username and API password moved from the connector profile to the linked account, so each RECRU account is connected with its own tenant, username and password. Replaced the setup guide with a single connection guide that covers finding the tenant subdomain, signing in, and copying the API credentials. Every list action now states its 50-record page limit, so agents page with offset instead of sending a limit the API rejects. Corrected descriptions that sent agents the wrong way: tags are named by label, country and locale codebooks return their real id shapes, get_project, get_candidate and get_department take a numeric id, work-position from, to and period cannot be sorted on, and education school cannot be searched. Preferred work locations now explain which place ids are accepted and where city ids come from, and the two CV actions say which one returns a readable PDF. Descriptions also state that projects hidden from the connected user return Record not found, that the offer list can lag behind recent edits because RECRU caches it, that a presentation stage can only be set in the RECRU backoffice, and that a candidate cannot be deleted through the API.
recru_create_candidate_preferred_work_locations, recru_create_project_candidate_phase_event, recru_get_candidate, recru_get_candidate_cv_content, recru_get_candidate_file_content, recru_get_department, recru_get_project, recru_list_candidate_educations, recru_list_candidate_events, recru_list_candidate_files, recru_list_candidate_preferred_locations, recru_list_candidate_tests_results, recru_list_candidate_work_positions, recru_list_candidates, recru_list_candidates_simple, recru_list_country_pairs, recru_list_department_contacts, recru_list_department_files, recru_list_departments, recru_list_departments_simple, recru_list_locales, recru_list_offers, recru_list_project_candidates, recru_list_projects, recru_list_projects_simple, recru_list_tag_sectors, recru_list_tag_seniorities, recru_list_tags, recru_list_users, recru_update_candidate
​
12 Oct 2026
new-actions data-sync webhooks docsAdded webhook support for job offer publish and unpublish events, delivered through RECRU’s Site API Listener. Documented the shape of the paging, search, specialFilter, profilePhoto, filesPreview and CV data objects, corrected action descriptions that named parameters the API does not accept, added list_project_close_reasons, and replaced account-specific code lists on project type, phases and close reason with pointers to the actions that resolve them. Allowed search values to be arrays for the in and notIn conditions, documented that notLike compares the value as given so a substring negation needs %value%, and documented the fields the provider truncates, overrides or does not store. Added validation so a request that would be silently ignored is rejected instead — jobseeker_source and source_name are now mutually exclusive, status_closed_reason is required when a project is closed, and the candidate tag actions require the candidate and tag ids. Added length limits on the department postal code and NACE fields so an over-long value is rejected instead of silently truncated, documented which fields the project-candidate list can be ordered by, recorded that a rejection phase requires a rejection reason, and removed the candidate status field, which the API accepts but never stores. Removing a candidate from a project now reports the real outcome — the provider commits the removal and then fails on a follow-up query, so the result is verified before being returned. Constrained the sortable columns on the project-candidate list, and removed the event id and the project-phase filter, which the API accepts but ignores. Documented that a project needs a type set before its pipeline can be used — without one, phase events and candidate removals fail with a database error. Corrected the file attachment fields — department files use isLogo, documentType now carries its real id list, and the stored name is the supplied filename with a unique prefix rather than a replacement.