API

Invoice OCR API

REST endpoint for invoice OCR — digital PDFs, scans, and phone photos. Returns the same typed JSON schema as the invoice extractor. Authenticate with a bearer API key created on the developers page.

Prefer the browser? Try the free invoice extractor. API access requires a paid plan.

Quick start

# Create a key at extractfox.com/developers, then:
curl -X POST https://extractfox.com/api/v1/extractions \
  -H "Authorization: Bearer efk_live_..." \
  -H "Content-Type: application/json" \
  -d '{"mode":"prebuilt","schema_id":"invoice","url":"https://example.com/invoice.pdf"}'

Example response

{
  "id": "job_9f3c1a7b2d4e8c05",
  "object": "extraction",
  "status": "succeeded",
  "mode": "prebuilt",
  "schema_id": "invoice",
  "source": { "kind": "url", "name": "invoice.pdf" },
  "result": {
    "vendor": "Acme Supplies Ltd.",
    "invoice_number": "INV-00284",
    "issue_date": "2026-04-12",
    "due_date": "2026-05-12",
    "currency": "USD",
    "line_items": [
      {
        "description": "A4 paper, 80gsm",
        "quantity": 12,
        "unit_price": 4.5,
        "amount": 54.0
      }
    ],
    "subtotal": 232.0,
    "tax": 46.4,
    "total": 278.4
  },
  "empty": false,
  "empty_fields": [],
  "error": null,
  "quota": { "used": 1, "limit": 300, "remaining": 299, "reason": "ok", "plan": "pro" }
}

Request fields

Field	Description
mode	prebuilt
schema_id	invoice
file	Object with base64 data, media_type, and name — PDF or image (PNG, JPEG, WEBP, HEIC), up to 3 MB inline. Scanned and photographed invoices supported.
url	Link to the invoice instead of inline data — fetched as a file (PDF up to 50 MB, image up to 20 MB).
async	Set true to get a 202 and poll GET /api/v1/extractions/{id} for the result.

Errors & limits

400Invalid JSON body — missing source, unknown schema_id, or more than one of file/url/text.
401Missing or revoked API key.
402Monthly extraction quota exhausted — upgrade or wait for reset.
413Inline file over 3 MB — host the document and pass its url instead (PDF up to 50 MB, image up to 20 MB).
415Unsupported media_type — use PDF or image.
429Rate limit — max 60 extraction calls/minute per IP.

Monthly extraction quotas apply per plan. See pricing and the general developers docs.

FAQ

Is this an invoice OCR service or an API?+

Both — the web invoice extractor is the no-code interface; this page documents the same extraction as a REST call for AP pipelines and ERP ingest.

Does the invoice OCR API work on scans and photos?+

Yes. Pass the same PDF or image you would upload in the browser. Multimodal extraction reads layout semantically — no per-supplier template training.

How do I authenticate?+

Create an API key on extractfox.com/developers and send it in the Authorization header as Bearer efk_live_… Keys are shown once, can be revoked there, and work for server-to-server calls and the MCP endpoint.

What does invoice OCR AI change vs classic OCR?+

Template OCR returns flat text per coordinate. Invoice OCR AI returns structured fields and line_items[] in one response — new supplier layouts work without retraining.

See pricing Try the web tool free →