OPENHIRE
For employers Protocol spec GitHub
● résumé never leaves your device
THE JOB DATA LAYER FOR AI AGENTS

Jobs come to your agent.
Your résumé never comes to us.

OpenHire is the open protocol behind AI-native hiring — verified listings scraped straight from employer career pages, queried by your own AI assistant. No account. No résumé upload. No ranking for sale.

$ claude mcp add openhire → Read the protocol spec
● résumé stays local ◆ ranking not for sale ▲ pay only for outcomes
FOR JOB SEEKERS

Stop uploading your résumé into a black hole.

▤ Matching happens on your machine Your skill fingerprint is generated locally. Only an anonymous vector ever reaches our server.
◈ It watches, so you don't have to Register a standing intent once. New matches ping you minutes after they go live on the employer's own site.
✓ You authorize every send One confirmation per job. It goes out under your name, through the employer's own channel — never a platform mailbox.
Walk through the flow →
FOR EMPLOYERS

Stop paying for exposure. Pay for outcomes.

Every application that reaches you has already read the JD and said yes — sent under the candidate's real identity, through your own career site or ATS. CPQC billing (pay only per authorized, qualified candidate) ships in v0.2. Claim your listings now for a verified badge and early access.

Notify me at v0.2 →
HOW IT'S BUILT

Three layers, one open protocol.

01 · DATA ENGINE Crawl employer career pages + public ATS APIs, verify freshness, score ghost-listing risk.
02 · PROTOCOL GATEWAY MCP · CLI · Skill — the same job graph, however your agent prefers to ask.
03 · TRUST & SETTLEMENT ghost_score, response SLA, CPQC — the layer that makes results, not exposure, the billable unit.
Full architecture in the pitch deck →
résumé stays local · ranking not for sale · pay only for outcomes
Protocol spec GitHub Pitch deck Live walkthrough