# Resume parsing and candidate matching, as an API and MCP server

> Source: <https://www.rezmatch.ai/>
> Published: 2026-07-29 18:01:22+00:00

# Parse any résumé.

Match any role.

The parse-and-match API for hiring products. Résumés and job descriptions in — clean JSON and calibrated, explainable fit scores out. Résumés are never stored.

No credit card. First call in under five minutes.

```
$ ▍
```

Built for the products that hire

## Start with any résumé

PDF, raw text, or a URL. Any industry, any format — nurses and accountants, not just engineers.

## Out comes clean, normalized JSON

E.164 phones, canonical titles and employers, computed tenures, taxonomy-pinned skills.

## Scored against the role, with receipts

Hard requirements gate the score. Every point is explained — evidence, never verdicts.

## The matcher is the product

Anyone can cosine-similarity two documents. Calibrated scores that hiring teams can defend — that's the hard part, and it's the part we obsess over.

### Parsing that survives real résumés

Any industry, any formatting — nurses and accountants, not just engineers. E.164 phones, canonical employers and titles, computed tenures, taxonomy-pinned skills.

### Scores with receipts

Met and missed requirements with evidence, and hard gates that similarity can't sweet-talk: a missing must-have caps the score, period.

### Fair by architecture

Every requirement verdict is grounded in quoted résumé evidence, and automated parity tests verify that name, age and ethnicity signals never move a score.

### Nothing to breach

Documents are deleted the moment your response is produced. No candidate database, no retention, no liability surface.

## One request.

Both documents. A defensible score.

Every endpoint takes text, url, or file — job-board URLs fetch directly, PDFs parse natively. Uniform envelope, stable error codes, failures refunded.

- ✓OpenAPI 3.0 spec + interactive playground
- ✓Parse once, match many — reuse candidate JSON across a whole pool
- ✓Free normalize endpoints to smoke-test your integration

```
curl -X POST https://api.rezmatch.ai/match \
  -H "x-access-key: $REZMATCH_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "resume_url": "https://files.acme.dev/resume.pdf",
    "jd_url": "https://job-boards.greenhouse.io/acme/jobs/123456"
  }'
```

## Your recruiting agent already knows how to use it

A hosted MCP server with OAuth connect, plus an Agent Skills pack for coding agents. The fairness rules ship inside both — an agent using Rezmatch.ai can't skip the parity-tested scoring path or turn scores into verdicts.

REST, MCP, and Skills are one surface: same credits, same limits, same explanations. No agent-only pricing games.

## Credits, not seats

A full screening — parse a résumé, match it against a role — is 8 credits. Failed calls are refunded. Per-credit price falls as tiers rise, and paid tiers accept larger documents at higher rate limits.

[250 creditsone-time · never expires$15](https://dashboard.rezmatch.ai)

[1,000 creditsone-time · never expires$55](https://dashboard.rezmatch.ai)

[3,000 creditsone-time · never expires$150](https://dashboard.rezmatch.ai)

Packs cost more per credit than any subscription — if you buy regularly, a plan is the better deal.