Metix AI Platform

Metix AI Platform is a data API and MCP server from OpenJobs AI Inc. that lets developers and AI agents query three datasets: people profiles, job postings and companies. It is reachable over REST, an MCP server and installable agent skills, all under one API key. The homepage states the index covers 900M profiles, 90M job postings and 30M companies.

What It Is

The Platform is a programmatic data layer for the job market, built for developers creating recruiting, sales or market-research tools, and for people who want agents such as Claude Code, Codex, Grok or Muse to look up people, postings and companies. It is separate from Metix AI, the company's AI recruiter product; the Platform returns data only and does not send outreach or book interviews.

How Search and Retrieval Work

Every dataset follows a two-step flow. A search, written either as a Query Spec tree or as a natural-language sentence for people, returns encrypted IDs and a cursor. The IDs, up to 100 per call, are then sent to the dataset's detail route to retrieve records, with fields selectable via _source. The same query grammar applies to people, jobs and companies, and results can be joined across datasets, for example to attach company context to a job. A separate contact capability returns work email, personal email or phone for people already found.

Agent Access via MCP and Skills

The MCP server is offered over streamable HTTP, with a legacy SSE endpoint, and exposes twelve tools: nine for data plus contract, version and key status. It uses the same key, limits and credit rules as REST. Four agent skills, covering people, company, job and cross-dataset search, install with a single npx command and read field names from a live contract endpoint. Setup guides exist for Claude, Codex and Muse.

Published Behavior

The docs state that queries refused for unknown fields or malformed trees, timeouts, upstream errors and empty Query Spec searches are not charged. No SDK is required; plain HTTP requests are used. A Casebook of reproducible example analyses documents the prompts, queries and data behind each case.