Greenhouse

This page records how composerID publishes into Greenhouse: the vendor’s API posture, the publish target and field mapping, and the documentation each claim rests on. It is for anyone evaluating the connection or building against it. Greenhouse is an ATS - the recruiting system of record on the permanent-hiring channel. composerID creates the Job from a tenant-designated template with its Openings, carries the Intent ID as the queryable requisition_id plus custom fields, and back-syncs offer and hire events through signed webhooks.

Category: ATS. Coverage status: Specification mapped. Vendor documentation: public developer portal.

Integration path

How composerID connects to this destination, at a glance.

Step	Detail
Authentication	Harvest API keys over HTTP Basic (key as username, blank password), created per-integration with per-endpoint permissions in Dev Center; every write requires an On-Behalf-Of header carrying an active Greenhouse user ID.
composerID publishes	Job: POST /v1/jobs (template_job_id)
Intent ID lands on	`job` `requisition_id`, written into the destination record
Back-sync	Webhooks: the destination pushes events; composerID verifies each delivery and lands it on the intent’s timeline.

Vendor documentation

Harvest API v3 (latest)

Harvest API reference (v1/v2)

Create or update a job via Harvest

Greenhouse API overview

API posture

Aspect	Detail
Authentication	Harvest API keys over HTTP Basic (key as username, blank password), created per-integration with per-endpoint permissions in Dev Center; every write requires an On-Behalf-Of header carrying an active Greenhouse user ID. Harvest v3 moves to OAuth 2.0 - v1/v2 are deprecated after 31 Aug 2026, so build against v3.
Style	REST, JSON (Harvest v1/v2 today, v3 latest); separate public Job Board API for published posts
Base URL	https://harvest.greenhouse.io/v1 (v2/v3 for select endpoints; v3 docs at harvestdocs.greenhouse.io)
Webhooks	UI-configured (Dev Center > Web Hooks); HMAC-SHA256 Signature header with a shared secret; events incl. job created/updated, offer created/approved, candidate hired.
Events posture	Webhooks
Rate limits	50 requests per 10 seconds per key; 429 with X-RateLimit-Limit/-Remaining headers (no Retry-After - back off client-side).
Objects	jobs, openings, job_posts, candidates, applications, offers, scorecards, users, departments, offices, custom_fields

Publish target and field mapping

Each object below pairs the canonical intent fields with the destination’s own fields and operations. The mapping is indicative until it is confirmed against a tenant at onboarding.

Job

Create the job from a template

Mapping	Detail
Canonical to destination	role_title, department, office, openings count; requisition_id = intent_id
Target	POST /v1/jobs (template_job_id)
Requirement	Required
Notes	requisition_id is queryable (GET /v1/jobs?requisition_id=…) - a natural Intent ID carrier

Openings

Headcount to fill

Mapping	Detail
Canonical to destination	opening_ids (per-seat), cost_center, budget
Target	POST /v1/jobs/{id}/openings
Requirement	Required
Notes	Opening custom fields carry per-seat data

Custom fields + hire back-sync

Correlate and reconcile

Mapping	Detail
Canonical to destination	intent_id/intent_version (custom fields); offer + hire events
Target	PATCH job custom_fields (immutable field key) + signed webhooks
Requirement	Required
Notes	candidate_hired / offer events update the intent timeline

Harvest keys are provisioned per integration with granular endpoint permissions - request only what publishing needs, and route writes On-Behalf-Of a real (integration system) user for audit attribution. Job creation copies a tenant-designated template job (scorecards and hiring teams cannot be set via API), and custom job field values need a follow-up PATCH. Sandbox and production have different object IDs and keys. Harvest v1/v2 sunset after 31 Aug 2026 - target v3 (OAuth 2.0).

Next steps

How composerID connects

The five ways composerID reaches a destination, and the minimum a destination must offer.

Publishing

Plan, preflight and publish: how a decision becomes a valid record in a destination.

All destinations

Every destination composerID documents, with its category and coverage status.

⌘I

Get started

Concepts

Guides

Destinations

Upstream connectors

API reference

Endpoints

Manuscript API (Triage)

Manuscript endpoints

Resources

Integration path

Vendor documentation

Harvest API v3 (latest)

Harvest API reference (v1/v2)

Create or update a job via Harvest

Greenhouse API overview

API posture

Publish target and field mapping

Next steps

How composerID connects

Publishing

All destinations

​Integration path

​Vendor documentation

Harvest API v3 (latest)

Harvest API reference (v1/v2)

Create or update a job via Harvest

Greenhouse API overview

​API posture

​Publish target and field mapping

​Next steps

How composerID connects

Publishing

All destinations

Integration path

Vendor documentation

API posture

Publish target and field mapping

Next steps