Greenhouse

Greenhouse

Greenhouse is a talent acquisition platform for recruiting, interviewing, and onboarding.

Auth and API

• Harvest v3 only — MCPBundles connects via OAuth to Greenhouse.
• Write requests require On-Behalf-Of with a Greenhouse user id that has permission for the action. The saved connection includes this user id; list users when you need to resolve one.
• List endpoints paginate with per_page and cursor (from the Link header rel="next").
• Candidate activity history lives on notes in v3. There is no separate activity-feed resource.

Capabilities

• Read: list or fetch by id across jobs, applications, candidates, pipeline, interviews, offers, org structure, custom fields, prospects, approvals, and compliance (EEOC/demographics).
• Upsert: omit id to create, pass id to update (candidates, applications, jobs, offers, openings, job posts, departments, offices, interviews, custom fields, and job notes).
• Create: POST-only endpoints (notes, attachments, candidate educations/employments/tags, job hiring managers and owners).
• Actions: move, hire, reject, unreject, and convert application to candidate via the application action tools. For moves, from_stage_id / to_stage_id are job interview stage ids from the job's interview-stage listing, not stage_id on application rows.
• Delete: remove candidates, applications, openings, attachments, interviews, tags, and custom field rows. Jobs are retired in the Greenhouse UI; Harvest v3 has no job removal endpoint.

Data model

• Jobs — requisitions with departments, offices, interview plans, and openings (headcount).
• Applications — candidate submissions to a job; link candidate_id + job_id on create.
• Candidates — people in the pipeline; educations, employments, tags, and notes attach here.
• Prospects — sourced talent in pools before or without an active application.
• Offers, approvals, org — offers, approval flows, departments, offices, users, job boards.

Typical workflow

1. List jobs and departments to anchor job_id and org structure.
2. List or fetch candidates and applications for the requisition; use notes for activity history.
3. For pipeline moves, read job interview stages first, then call the application action with stage ids from that listing.