Building a Lightweight ATS with ParseApi: Resume Parsing Without the Lock-in

Most Applicant Tracking Systems cost thousands of dollars a month and lock you into their ecosystem. But for a team that needs structured data extracted from resumes — to route candidates, filter by skill, or feed a downstream system — the full ATS is overkill.

Here's how to build a lightweight resume parsing layer using ParseApi in under 30 minutes, with a REST API you own.

What you'll build

• A ParseApi folder that accepts resume PDFs
• An auto-detected schema covering the key candidate fields
• A queryable REST API you can filter by skill, experience, or location
• An optional webhook that fires whenever a new resume is processed

Step 1: Create the folder

Sign in to ParseApi and create a new folder:

• Slug: candidates
• Auth mode: API key (generate a live key for your system)

Step 2: Upload sample resumes

Upload 3–5 diverse resume PDFs. ParseApi will auto-detect a schema after the third upload. After the synthesis, you'll see a proposed schema along these lines:

{
  "full_name": "string",
  "email": "string",
  "phone": "string",
  "location": "string",
  "current_title": "string",
  "total_experience_years": "number",
  "skills": ["string"],
  "education": [
    {
      "institution": "string",
      "degree": "string",
      "graduation_year": "number"
    }
  ],
  "work_experience": [
    {
      "company": "string",
      "title": "string",
      "start_date": "string",
      "end_date": "string",
      "description": "string"
    }
  ]
}

Review the proposed schema, add any custom fields you need (e.g. cover_letter_summary, github_url, portfolio_url), and confirm.

Step 3: Query the candidate list

With a few resumes uploaded, query the folder:

curl https://api.parseapi.dev/v1/yourname/candidates \
  -H "Authorization: Bearer pk_live_..."

Each result includes the full extracted profile, the source document reference, and the extraction timestamp.

Step 4: Filter by skill

Use the field-search endpoint to find candidates with a specific skill:

curl "https://api.parseapi.dev/v1/yourname/candidates/search?field=skills&value=Python" \
  -H "Authorization: Bearer pk_live_..."

Find senior candidates by experience:

curl "https://api.parseapi.dev/v1/yourname/candidates/search?field=total_experience_years&value=gte:5" \
  -H "Authorization: Bearer pk_live_..."

Or filter by location:

curl "https://api.parseapi.dev/v1/yourname/candidates/search?field=location&value=New+York" \
  -H "Authorization: Bearer pk_live_..."

Step 5: Set up a webhook for new applications

When a candidate applies and a new resume lands in your folder, you want your system to react immediately. Set up a webhook in folder settings pointing to your endpoint:

https://yoursystem.example.com/api/new-candidate

ParseApi sends a POST request with the extraction result as soon as processing completes:

{
  "event": "extraction.completed",
  "folder": "candidates",
  "document_id": "01HZB2VQ9S4KFWMEC3GRX7YNJ4",
  "extracted_at": "2026-04-28T14:33:07Z",
  "result": {
    "full_name": "Sarah Chen",
    "email": "sarah@example.com",
    "location": "San Francisco, CA",
    "current_title": "Senior Backend Engineer",
    "total_experience_years": 6,
    "skills": ["Python", "PostgreSQL", "FastAPI", "Docker", "Kubernetes"]
  }
}

Your endpoint can then route the candidate to the right reviewer, trigger a Slack notification, send an auto-acknowledgment email, or update your internal pipeline tracker.

Handling multi-page and complex layouts

ParseApi processes every page of a multi-page PDF. A 3-page resume with a cover letter and reference section is extracted as a single result. Two-column layouts and table-formatted experience sections are handled well by the vision-capable AI models ParseApi routes to by default.

The page count is tracked in usage metering — a 3-page resume costs 3 pages from your monthly allocation.

When extraction gets something wrong

Resumes are notoriously variable. Unusual fonts, creative layouts, and overlapping columns all challenge AI extraction. When a field is wrong:

1. Open the extraction in the ParseApi dashboard
2. Click the incorrect field — the source PDF highlights where the value was found
3. Edit the value inline
4. The correction is saved immediately and reflected in the API

Corrections are tracked per-field, so over time you can see which resume formats are causing the most errors — useful for understanding which source (job boards, direct applications, agency submissions) has the cleanest documents.

Cost estimate

On the Hobby plan ($19/month) you get 1,000 pages. A typical resume is 1–2 pages, so that covers 500–1,000 candidate documents per month — plenty for a growing startup or a small recruiting agency.

The free tier (100 pages) is enough to evaluate the approach before committing. No credit card required.