Endpoint

Parse resume

Extract structured fields from a PDF, DOCX, or plain-text resume. Vision fallback for scanned image-only PDFs.

Authentication

Requires the parse scope.

Endpoint

POST/api/v1/parse/resume

Request — multipart upload

Send file as a multipart form field. PDF, DOCX, and plain text are all accepted. PDFs with no extractable text fall through to Claude's PDF-vision pipeline automatically.

curl -X POST https://www.astrishr.com/api/v1/parse/resume \
  -H "Authorization: Bearer $KEY" \
  -F "file=@candidate-resume.pdf"

Request — JSON text

Or pass the resume body as plain text:

textstringrequired

The full resume body. Astris truncates after 30K characters.

curl -X POST https://www.astrishr.com/api/v1/parse/resume \
  -H "Authorization: Bearer $KEY" \
  -H "Content-Type: application/json" \
  -d '{"text":"Mohammed K.\nAgricultural Supervisor — ACME Farm, 2018–2024\nKandahar, Afghanistan\nSkills: team leadership, scheduling, tractor, inventory\nLanguages: Dari, Pashto, English"}'

Response

{
  "parsed": {
    "skills": ["Team Leadership", "Scheduling", "Tractor Operation", "Inventory Management"],
    "certifications": [],
    "languages": ["Dari", "Pashto", "English"],
    "yearsExperience": 6,
    "summary": "Agricultural supervisor with 6 years' experience...",
    "workHistory": [
      { "employerName": "ACME Farm", "jobTitle": "Agricultural Supervisor", "startDate": "2018", "endDate": "2024", "country": "Afghanistan" }
    ],
    "education": []
  },
  "usedVision": false,
  "parseError": null
}

Notes

usedVision is true when Astris fell back to Claude vision (scanned PDF, < 50 chars extracted).
Reviewer corrections from DocumentParseRating are auto-injected as in-context examples to improve subsequent parses.
Foreign credentials and former occupations are preserved exactly as written — they drive transferable-skill matching.