Skip to Content

Parse a resume to JSON

Turn a resume into clean JSON with contact details, work history, education and skills. Each contact fact comes with the line of the resume it came from.

Your users upload a resume. Your product needs fields: a name, an email, past jobs, skills. POST /developer/parse turns resume text into that JSON in one call. Each contact fact comes with the exact resume line behind it. So your product fills a form with what the person wrote, never with a guess.

You need an API key. The Quickstart shows how to get one.

1. Get the text out of the file

The API reads text: plain text or Markdown, 50 to 50,000 characters. For a PDF, pull the text out on your server first. In Node.js or Bun, the open source unpdf package does it in one call:

pdf-to-text.ts
import { extractText } from 'unpdf';

export async function pdfToText(file: ArrayBuffer) {
  const { text } = await extractText(new Uint8Array(file), { mergePages: true });
  return text;
}

A scanned PDF holds a picture of the page and no text layer, so it gives back little or no text. Check the length before you call the API. Text under 50 characters gets a 422. See Errors.

2. Send the text

curl
curl https://api.getmattered.com/api/v1/developer/parse \
  -H "Authorization: Bearer $GETMATTERED_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"resumeText": "Maya Patel\nSenior Data Analyst\nAustin, TX | maya.patel@example.com | +1 512 555 0142\nlinkedin.com/in/mayapatel-example\n\nExperience\nSenior Data Analyst, Northwind Health, Mar 2021 - Present\nBuilt SQL and Python pipelines for weekly reporting.\nData Analyst, Contoso Retail, Jun 2017 - Feb 2021\nRan A/B testing for pricing changes.\n\nEducation\nB.S. Statistics, University of Texas at Austin, 2017\n\nSkills\nSQL, Python, Tableau, Excel, A/B testing"}'
TypeScript (fetch)
const response = await fetch('https://api.getmattered.com/api/v1/developer/parse', {
  method: 'POST',
  headers: {
    'Authorization': `Bearer ${process.env.GETMATTERED_API_KEY}`,
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({ resumeText: await pdfToText(file) }),
});

const body = await response.json();
if (!body.success) {
  throw new Error(`${body.error.code}: ${body.error.message}`);
}
const { data } = body;
console.log(data.fullName, data.email, data.workExperience.length);

3. Read the response

A successful call answers 200. This is the response for the resume in step 2, trimmed for space:

POST /developer/parse 200 response
{
  "success": true,
  "data": {
    "fullName": "Maya Patel",
    "firstName": "Maya",
    "lastName": "Patel",
    "email": "maya.patel@example.com",
    "phone": "+1 512 555 0142",
    "city": "Austin",
    "state": "TX",
    "country": null,
    "zipCode": null,
    "linkedinUrl": "linkedin.com/in/mayapatel-example",
    "githubUrl": null,
    "portfolioUrl": null,
    "currentJobTitle": "Senior Data Analyst",
    "professionalSummary": null,
    "workExperience": [
      {
        "title": "Senior Data Analyst",
        "company": "Northwind Health",
        "startDate": "Mar 2021",
        "endDate": null,
        "current": true,
        "description": "Built SQL and Python pipelines for weekly reporting."
      },
      {
        "title": "Data Analyst",
        "company": "Contoso Retail",
        "startDate": "Jun 2017",
        "endDate": "Feb 2021",
        "current": false,
        "description": "Ran A/B testing for pricing changes."
      }
    ],
    "education": [
      { "degree": "B.S. Statistics", "institution": "University of Texas at Austin", "year": "2017" }
    ],
    "skills": ["SQL", "Python", "Tableau", "Excel", "A/B testing"],
    "fieldEvidence": [
      { "field": "city", "quote": "Austin, TX | maya.patel@example.com | +1 512 555 0142", "confidence": "high" },
      { "field": "email", "quote": "Austin, TX | maya.patel@example.com | +1 512 555 0142", "confidence": "high" },
      { "field": "phone", "quote": "Austin, TX | maya.patel@example.com | +1 512 555 0142", "confidence": "high" },
      { "field": "linkedinUrl", "quote": "linkedin.com/in/mayapatel-example", "confidence": "high" },
      { "field": "currentEmployer", "quote": "Senior Data Analyst, Northwind Health, Mar 2021 - Present", "confidence": "high" },
      { "field": "educationLevel", "quote": "B.S. Statistics, University of Texas at Austin, 2017", "confidence": "low" }
    ],
    "experienceLevel": "senior",
    "educationLevel": "bachelor",
    "yearsOfExperience": 9,
    "industries": ["Healthcare", "Retail"],
    "jobTypes": ["full_time"],
    "companySize": [],
    "suggestedJobTitles": ["Senior Data Analyst", "Analytics Engineer"],
    "suggestedLocations": ["Austin, TX", "Remote"],
    "usage": { "tokens": 530 }
  },
  "meta": { "requestId": "...", "timestamp": "2026-11-18T12:00:00.000Z", "version": "v1" }
}

The fields fall into three groups:

GroupFieldsWhat to do with them
Read from the resumefullName, email, phone, city, linkedinUrl, currentJobTitle, workExperience, education, skills and the other contact fieldsUse them. A fact the resume does not print is null or an empty list.
Inferred from the whole resumeexperienceLevel, educationLevel, yearsOfExperience, industries, jobTypes, companySize, suggestedJobTitles, suggestedLocationsShow them as suggestions the person can change.
EvidencefieldEvidenceCheck a contact fact before you fill a form with it. See step 4.

Dates stay as the resume prints them, such as Mar 2021. The API reference lists every field, including the ones this example leaves out.

4. Fill a form only with what the resume says

Each fieldEvidence entry names a field, quotes the resume text that holds its value, and rates it. high and medium mean the value is printed in the resume. low means it was inferred.

Prefill a contact field only when the resume prints it:

printed-fields.ts
type Evidence = { field: string; quote: string | null; confidence: 'high' | 'medium' | 'low' };

export function printedFields(evidence: Evidence[]) {
  return new Set(
    evidence
      .filter(item => item.confidence !== 'low' && item.quote !== null)
      .map(item => item.field),
  );
}

const printed = printedFields(data.fieldEvidence ?? []);
const prefill = {
  email: printed.has('email') ? data.email : null,
  phone: printed.has('phone') ? data.phone : null,
  city: printed.has('city') ? data.city : null,
  linkedinUrl: printed.has('linkedinUrl') ? data.linkedinUrl : null,
};

Show the quote next to the field in your form. The person sees where each value came from and fixes it in one tap.

What it costs

Each parse spends API tokens from the account that owns the key. data.usage.tokens says how many this call spent. Pricing has the usual cost of every call.

Next

On this page