Open Job Protocol

Why Open Job Protocol?

Job postings today are unstructured prose optimized for SEO, not for AI agents. schema.org/JobPosting serves Google, not hiring agents. Every ATS, job board, and company career page uses a different format. Critical signals like salary, remote policy, and team composition are buried in paragraphs or missing entirely.

Open Job Protocol (OJP) fixes this. v0.2 restructures postings into flat, snake_case top-level fields with strict additionalProperties: false, direct alignment with OTP v0.2 field names, and ESCO taxonomy support for skills. OJP is the employer half of a bilateral ecosystem — its companion standard, the Open Talent Protocol, handles the candidate and profile side.

Goal	What it means in practice
Machine-first	JSON is the canonical format. Flat top-level fields, snake_case keys, and strict `additionalProperties: false` for reliable parsing.
Employer-controlled	`salary_transparency` lets employers decide when to disclose compensation. No data is shared without explicit intent.
Agent-ready	`must_have` and `nice_to_have` are structured objects agents can filter and rank on — not prose to be parsed.
OTP-aligned	Field names map directly to OTP v0.2 — `salary_band`, `seniority`, skill `level`, and CEFR language proficiencies match exactly.
Vendor-neutral	MIT licensed. No required registry, platform, or ATS.
Interoperable	Mappable to schema.org/JobPosting, Google for Jobs, LinkedIn Job Posting API, and HR-XML. Full v0.1 migration guide included.

Quick start

Validate a document

The validator CLI checks any JSON file against the schema and prints readable errors.

# Install and build the validator
git clone https://github.com/neogene-ai/open-job-protocol.git
cd open-job-protocol/tools/validator-cli
npm install && npm run build

# Validate one of the v0.2 example documents
node dist/index.js ../../examples/v0.2/backend-engineer-senior.json
✓ Valid Open Job Protocol document

Minimal valid document

A valid v0.2 document needs only the 10 required top-level fields. Everything else — salary, requirements, team, process — is optional.

{
  "schema_version": "0.2.0",
  "ojp_id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
  "created_at": "2026-03-30T10:00:00Z",
  "updated_at": "2026-03-30T10:00:00Z",
  "status": "active",
  "title": "Senior Frontend Engineer",
  "description": "Build and maintain our React-based product...",
  "employment_type": "full_time",
  "organization": { "name": "Acme Corp" },
  "location": { "arrangement": "hybrid" }
}

Schema reference

The full schema is a JSON Schema (draft 2020-12) document at schema/v0.2/ojp.schema.json. Below, each section is documented with its fields, types, and intent.

Looking for v0.1? The v0.1 schema remains available. See the v0.1 → v0.2 migration guide for a full field mapping.

Top-level fields

required

Every OJP v0.2 document must include these 10 required fields.

Property	Type	Req.	Description
schema_version	string	yes	Must be `"0.2.0"`.
ojp_id	uuid	yes	Unique job posting identifier.
created_at	date-time	yes	Posting creation timestamp (ISO 8601).
updated_at	date-time	yes	Last modification timestamp (ISO 8601).
status	enum	yes	Lifecycle status: draftactivepaused closedexpiredfilled
title	string	yes	Job title (max 200).
description	string	yes	Full role description (max 5000).
employment_type	enum	yes	full_timepart_timecontract freelanceinternshiptemporary
organization	object	yes	Hiring entity. See organization.
location	object	yes	Work location and arrangement. See location.

Optional top-level fields

Property	Type	Description
valid_through	date	Posting expiration date.
summary	string	One-paragraph agent-optimized role summary (max 500).
seniority	enum	Matches OTP `seniority` exactly: internjuniormid seniorstafflead principaldirectorvpc_level
function	string	Organizational function (max 100). E.g. `"Engineering"`, `"Sales"`.
total_openings	integer	Number of open positions (min 1).
salary_transparency	enum	When salary is disclosed: publicon_requestafter_interview

organization

required

The hiring entity. Only name is required within this object.

Property	Type	Req.	Description
name	string	yes	Company name (max 200).
url	uri	no	Company website.
industry	string	no	Primary industry (max 100).
size	enum	no	Matches OTP `preferences.company_size`: startupscale_upmid_marketenterprise
department	string	no	Hiring department (max 100).
headquarters	string	no	HQ location (max 200).

location

required

Work location and arrangement. Only arrangement is required.

Property	Type	Req.	Description
arrangement	enum	yes	Matches OTP `work_model`: onsitehybridremote
country	string	no	ISO 3166-1 alpha-2. Matches OTP `location.country`.
city	string	no	City name (max 100).
region	string	no	State or region (max 100).
remote_regions	string[]	no	ISO 3166-1 alpha-2 country codes for remote eligibility.
relocation_support	boolean	no	Whether relocation assistance is offered.
visa_sponsorship	boolean	no	Whether visa sponsorship is available.

salary_band

optional

Compensation range. Same structure as OTP salary_band for bilateral matching.

Property	Type	Req.	Description
min	number	yes	Minimum salary.
max	number	yes	Maximum salary.
currency	string	yes	ISO 4217 currency code (e.g. `"EUR"`, `"USD"`).
period	enum	yes	annualmonthlydailyhourly

must_have

optional agent-critical

Hard requirements — candidates must meet all of these. An OJP-compliant agent must not present a candidate who fails any must_have constraint.

Property	Type	Description
skills	skill_requirement[]	Each: `name` (required), `min_level` (1–5, maps to OTP `skills[].level`), `min_years`, `esco_id`.
experience_years	object	`min` and optional `max`.
credentials	string[]	Required degrees or qualifications (max 200 each).
certifications	string[]	Required certifications (max 200 each). E.g. `"CKA"`, `"ACLS"`.
languages	language_requirement[]	Each: `language` (ISO 639-1), `proficiency` (CEFR): A1A2B1B2C1C2native Matches OTP `languages[].proficiency` exactly.
work_authorization	string[]	ISO 3166-1 alpha-2 country codes. Matches OTP `visa_status.authorized_countries`.

nice_to_have

optional

Preferred qualifications — bonus scoring, not hard filters.

Property	Type	Description
skills	skill_requirement[]	Same shape as `must_have.skills`.
experience_years	object	`preferred` years.
credentials	string[]	Preferred qualifications.
certifications	string[]	Preferred certifications.
languages	language_requirement[]	Same shape as `must_have.languages`.

Design note: The must_have / nice_to_have split is the key agent differentiator. An OJP-compliant agent must treat must_have as a hard filter and nice_to_have as ranking signal. This is a hard rule, not a hint.

benefits

optional

Array of benefits offered by the employer.

Property	Type	Req.	Description
category	enum	yes	healthretirementdevelopment flexibilityfamilyfinancialother
description	string	yes	Benefit details (max 500).

team

optional

Team composition and context for culture matching.

Property	Type	Description
name	string	Team or squad name (max 100).
size	integer	Current team size.
reports_to	string	Hiring manager title (max 200).
tech_stack	string[]	Team's technology stack (max 100 each).
methodology	string	Work methodology (max 100). E.g. `"scrum"`, `"kanban"`.
description	string	Team mission and purpose (max 1000).

process

optional

Hiring process transparency — stages, timeline, and application method.

Property	Type	Description
stages	hiring_stage[]	Each: `name`, `type`: screeningtechnicalbehavioral case_studytake_homeofferother and optional `duration_minutes`.
total_duration_days	integer	Target time-to-offer in days.
decision_timeline	string	When candidates hear back (max 200).
application_url	uri	Where to apply.
accepts_otp_profile	boolean	Whether OTP profiles are accepted as applications.
ai_screening	boolean	Transparency: is AI used in candidate screening?

culture

optional

Culture signals for work-style and values matching.

Property	Type	Description
values	string[]	Core company values (max 100 each).
work_style	string[]	Work style descriptors (max 100 each). E.g. `"async-first"`, `"low-meeting"`.

source

optional

Data provenance — which agent or system created this posting.

Property	Type	Req.	Description
agent_id	string	yes	Identifier of the agent or system (max 200).
platform	string	no	Platform name (max 200). E.g. `"lever"`, `"greenhouse"`.

OTP ↔ OJP bilateral matching

Symmetric matching between OTP v0.2 candidate profiles and OJP v0.2 job listings. An agent with both documents can compute a bilateral match score without any natural language processing.

OTP v0.2 (Candidate)	OJP v0.2 (Job)	Match type	Logic
skills[].name / .level	must_have.skills[].name / .min_level	hard filter	Skill name match + level ≥ min_level
skills[]	nice_to_have.skills[]	soft rank	Bonus score for matching preferred skills
certifications[]	must_have.certifications[]	hard filter	Certification name match
experience[].duration	must_have.experience_years.min	hard filter	Sum of durations ≥ min
languages[].proficiency	must_have.languages[].proficiency	hard filter	CEFR level ≥ required level
salary_band	salary_band	bilateral	candidate.min ≤ job.max AND job.min ≤ candidate.max
work_model	location.arrangement	hard filter	Arrangement in candidate's work_model array
visa_status.authorized_countries	must_have.work_authorization	hard filter	Candidate authorized in required country
seniority	seniority	soft rank	Exact or adjacent match
preferences.culture_values	culture.values	soft rank	Overlap scoring

Discovery

Employers host a well-known discovery file so agents can find all active listings without scraping.

# https://acme.com/.well-known/ojp.json

{
  "version": "0.2.0",
  "organization": "Acme Corp",
  "jobs_url": "https://acme.com/api/jobs.ojp.json",
  "total_active_jobs": 12,
  "updated_at": "2026-03-28T14:00:00Z",
  "accepts_otp_profile": true,
  "contact": "careers@acme.com"
}

Examples

Three schema-conforming v0.2 documents are included in the repository. They demonstrate how must_have/nice_to_have, salary_band, and process are used across different industries and seniority levels.

Engineering

Senior Backend Engineer

Hybrid Go/Rust role in Berlin. Full process stages, salary band, ESCO skills, and team details.

Product

Product Manager

Remote-first B2B SaaS role. Culture signals, growth path, and structured case-study hiring process.

Healthcare

Nurse Practitioner

Onsite clinical role with required certifications, language requirements, and shift differentials.

Tools

Validator CLI

A Node.js CLI that validates any JSON file against the OJP schema and prints readable errors. Source: tools/validator-cli/

cd tools/validator-cli
npm install && npm run build
node dist/index.js path/to/job.json

Agent tools

TypeScript stubs for integrating OJP into MCP-compatible agent frameworks. Source: tools/agent-examples/

Provides composable functions ready to be registered as MCP tools: validateJobPosting, matchCandidateToJob, and introspectJobPosting.

MCP server

A standalone MCP server exposing all 6 OTP/OJP agent tools (validate, introspect, parse for both schemas). Source: tools/mcp-server/ (hosted in the OTP repo).

Migration

OJP is designed to be clearly mappable to and from existing job posting formats.

Format	Guide	Notes
OJP v0.1	v0.2-migration.md	Full field mapping from nested (meta, role, offering) to flat top-level structure. v0.1 schema remains available.
schema.org/JobPosting	migration-schemaorg.md	Full field-by-field mapping. OJP is a superset of schema.org’s job properties.
HR-XML	migration-hrxml.md	Conceptual guide. HR-XML’s PositionOpening maps to OJP without data loss.
LinkedIn Job Posting API	migration-linkedin.md	Maps LinkedIn’s foundation schema fields to OJP sections.

Ecosystem

OJP defines the employer side of agentic marketplaces. It is designed to work alongside complementary open standards that together cover both sides of the hiring equation.

Open Talent Protocol (OTP) is the candidate-side counterpart to OJP — a machine-first, vendor-neutral JSON standard for structuring professional profiles. OTP v0.2 uses progressive disclosure tiers to control how much data is revealed at each stage of matching, so 80% of agent traffic can run on the cheapest models.

opentalentprotocol.org · GitHub repository

Together, OTP + OJP enable fully structured, agent-driven bilateral matching. Employers express requirements and offer details in an OJP posting; candidates express preferences and constraints in their OTP profile. Agents on both sides can evaluate fit and negotiate programmatically — no unstructured job-board listings or PDF resumes required.

Governance & neutrality

Open Job Protocol is vendor-neutral. JobGrow and neogene.ai are its initial sponsors, but the spec is designed for broad ecosystem adoption. No single company has veto power over the schema.

OJP is the employer-side counterpart to Open Talent Protocol. Together, they form a complete, open standard for agentic talent marketplaces. The same community develops both protocols.

The project is governed by a maintainer council using lazy consensus. See GOVERNANCE.md for the full process.

Licensed under MIT. The spec will always be free to implement.