Skip to Content
Introduction

Rezmatch.ai

Parse and match, as an API. Rezmatch.ai turns résumés and job descriptions into clean, normalized JSON and scores candidate-role fit with calibrated, explainable results — so hiring products can build screening features without building an extraction pipeline.

curl -X POST "$BASE/parse/resume" \
  -H "x-access-key: $REZMATCH_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://files.example.com/resume.pdf"}'

What it does

  • Parse — résumés and JDs → structured JSON with consistent formats: E.164 phones, canonical locations, normalized employers/titles/institutions, computed tenures, taxonomy-pinned skills. Works for any industry, not just tech.
  • Match — one candidate against one role → a 0-100 score with a band, met/missed requirements with evidence, and grounded prose notes. Hard requirements dominate; text similarity can’t rescue a missing must-have.
  • Extract & normalize — à-la-carte: skills, experience, education, contact, plus free title/skill normalization endpoints.
  • Redact — strip PII and protected-class proxies for bias-free review.

Three ways to integrate

  1. REST API — every endpoint accepts GET or POST, inputs as text, url, or file (base64 PDF). Start with the quickstart.
  2. MCP connector — give Claude (or any MCP client) the tools directly. Paste one URL, sign in, done. See MCP Connector.
  3. Agent Skills — drop-in skill files that teach coding agents the safe workflows. See Agent Skills.

Two promises

  • Résumés are never stored. Processing is transient; documents are deleted immediately after the response is produced. There is no candidate database to breach.
  • Scores and explanations, never verdicts. Matching always runs on a bias-redacted feature view (fairness.scored_on_redacted_input is always true), and results are evidence for human decisions — the API will never say “hire” or “reject”.

Explore

Last updated on
Quickstart