GDPR-first document AI

Turn ID photos into structured data with one API call

Passports, identity cards and driving licences — parsed into clean JSON, MRZ check-digit validated, front and back merged automatically. Document contents are processed in memory and never stored.

Start free — 25 parses Read the docs

No credit card required.

POST /api/v1/parse

$ curl -X POST https://goparse.it/api/v1/parse \
    -H "Authorization: Bearer pid_live_************" \
    -F "images=@id_front.jpg" \
    -F "images=@id_back.jpg"

{
  "request_id": "req_7f3a2c9b1e",
  "persons": [
    {
      "family_name": "ROSSI",
      "given_name": "MARIA",
      "date_of_birth": "1991-04-12",
      "sex": "F",
      "citizenship": "ITA",
      "place_of_birth": "BOLOGNA (BO)",
      "document_type": "identity_card",
      "document_number": "CA12345AA",
      "issuing_authority": "MINISTERO DELL'INTERNO"
    }
  ],
  "warnings": [],
  "processing": {
    "images_received": 2,
    "documents_detected": 1,
    "elapsed_s": 23.4
  }
}

0documents stored — contents are processed in memory only

ICAO 9303check-digit validation on every machine-readable zone

EU-hostedprocessing on EU infrastructure, GDPR-first by design

Built for identity workflows, not photo albums

A research-grade extraction pipeline behind a boring, dependable REST API.

MRZ check-digit validation

Every machine-readable zone is verified against ICAO 9303 check digits, so you know when a field is confirmed by the document itself — not just read from it.

Front and back auto-merging

Send both sides of a card in one request. The pipeline detects which images belong to the same document and merges them into a single person record.

Multi-person, multi-document

A photo containing several documents — or a request covering several people — comes back as a clean persons array, one entry per person.

Rotation and orientation robust

Skewed, rotated or upside-down photos are detected and re-oriented automatically before extraction. No pre-processing needed on your side.

Empty over wrong

When the pipeline is not confident about a field, it returns an empty string — never a guess. Predictable behaviour your compliance team can rely on.

GDPR-safe usage dashboard

Track volume, latency and document types over time. Usage logs contain aggregate metadata only — never names, numbers or images.

How it works

Upload images

POST up to 8 JPEG or PNG images (15 MB each) to /api/v1/parse with your API key. Plain multipart — no SDK required.

Pipeline reads and validates

Documents are segmented, re-oriented and read by a vision-language pipeline; MRZ lines are verified against ICAO 9303 check digits.

Structured JSON

You receive a persons array with nine normalised fields per person, plus warnings and processing metadata. Typical turnaround: under 2 seconds.

Built for GDPR from day one

Identity documents are some of the most sensitive data your product will ever touch. We designed GoParse so that holding on to them is impossible, not just avoided.

In-memory processing
Images and extracted fields exist only for the lifetime of the request, then they are gone.
No document retention
Nothing is written to disk, nothing is cached, and your data is never used for training.
Metadata-only usage logs
We log document type, timestamps and counts for billing — never names, numbers or images.
DPA included
A standard Article 28 DPA is part of the terms on every plan; Enterprise adds a custom, countersigned DPA.

Read the DPA Privacy policy

Pricing that scales with you

Start free, upgrade when you go live.

Full pricing

Free Trial

Evaluate the API on your own documents

€0

25 parses included

Start for free

Starter

For small teams going live

€15 / month

500 parses / month

Choose Starter

Pro

Enterprise

Pay as you go, at your scale

Custom

Unlimited parses, pay as you go

Frequently asked questions

What happens to the images I upload?

They are processed entirely in memory for the duration of the request and discarded the moment your response is sent. Nothing is written to disk and nothing is retained. The only records we keep are generic usage metadata: document type, timestamps and counts — enough to show your dashboard and bill correctly, and nothing more.

Which documents are supported?

Passports, national identity cards and driving licences. Front and back images of the same card are merged automatically, and a single photo may contain several documents. Each result includes a document_type field telling you what was detected.

How accurate is it?

MRZ-backed fields are validated with ICAO 9303 check digits, so they are confirmed by the document itself. For everything else we follow a strict empty-over-wrong policy: if the pipeline is not confident about a field, it returns an empty string instead of guessing. That makes results predictable — you always know when a value needs human review.

How long does a request take?

Up to 2 seconds per request, depending on image count and quality — and that includes full segmentation, orientation correction, extraction and validation, not just a quick OCR pass. Each account can run 2 requests concurrently.

Can I test it without a credit card?

Yes. The free trial includes 25 document parses with all features enabled — no credit card required. Sign up, create an API key and send your first request in minutes.

Is there an SLA or an on-premise option?

Both are available on the Enterprise plan, which adds pay-as-you-go volume pricing, dedicated capacity and a custom DPA. Contact us to scope your deployment.

Parse your first document today

25 free parses, every feature enabled, nothing stored. If it does not fit your workflow, you have lost nothing — not even your data.

Get started for free Contact us