CV-Pooling API

1. Your company must be on a plan with API access included (Enterprise — contact sales if you're not sure).
2. In the dashboard, go to Settings → Account → API Access and click Generate new key.
3. Copy the key immediately — it's shown once and can't be retrieved again. If you lose it, revoke it and generate a new one.

Keep the key secret: it authenticates as your company and its use is billed against your account's credits.

Send the key on every request as the X-API-Key header:

X-API-Key: oxa_live_xxxxxxxxxxxxxxxxxxxxxxxx

POST /api/v1/cv-pooling/match

POST /api/v1/cv-pooling/match
Content-Type: application/json
X-API-Key: oxa_live_xxxxxxxxxxxxxxxxxxxxxxxx

{
  "job_title": "Senior Backend Engineer",
  "job_description": "We're looking for an engineer with 5+ years of Node.js, PostgreSQL, and AWS experience...",
  "candidates": [
    {
      "candidate_ref": "your-ats-id-123",
      "candidate_name": "Jane Doe",
      "cv_file_name": "jane-doe.pdf",
      "cv_url": "https://your-storage.example.com/cvs/jane-doe.pdf"
    },
    {
      "candidate_ref": "your-ats-id-124",
      "cv_file_name": "resume.pdf",
      "cv_base64": "JVBERi0xLjQK..."
    }
  ]
}

• job_title / job_description — required. The CV is matched directly against these — nothing needs to exist in Orixa first.
• candidates — 1 to 25 per request. Each candidate needs exactly one of:
  • cv_url — a publicly reachable URL to the CV (preferred for larger files)
  • cv_base64 — the raw file, base64-encoded (for CVs without a public URL)
• Supported file types: PDF and DOCX. Legacy .doc isn't supported.
• candidate_ref is your own identifier — echoed back on the result so you can match it up on your side (these CVs aren't stored in Orixa, so there's no internal id to return).

{
  "success": true,
  "message": "CVs matched successfully. 1 credit used, 42 remaining.",
  "data": {
    "job_title": "Senior Backend Engineer",
    "results": [
      {
        "candidate_ref": "your-ats-id-123",
        "candidate_name": "Jane Doe",
        "match_percentage": 87,
        "matched_skills": ["Node.js", "PostgreSQL", "AWS"],
        "missing_skills": ["Kubernetes"],
        "summary": "Strong backend experience with directly relevant stack...",
        "recommendation": "strong_fit"
      }
    ],
    "credits_used": 1,
    "credits_remaining": 42
  }
}

If a specific CV couldn't be scored (unreachable URL, unsupported format, malformed AI response), that candidate's result carries an error field instead of scores — it doesn't fail the rest of the batch.

Same pricing as the dashboard feature: 1 credit per 5 CVs matched, rounded up, deducted from your subscription before matching runs.