Extract data. Trace sources. Trust with confidence.

Turn PDFs and images into JSON. Check extracted values against the source page.

Open the playgroundExplore lineage
MUTUAL NON-DISCLOSURE AGREEMENT
This Agreement is made on 2 September 2026 between Nimbus Cloud GmbH, 42 Speicherstrasse, Berlin ("Nimbus") and Acme Operations GmbH, 8 Marktgasse, Vienna ("Acme").
4. Term and termination
This Agreement remains in force for three (3) years from the date above. Either party may end it with thirty (30) days' written notice.
5. Governing law
This Agreement is governed by Austrian law. The courts of Vienna have exclusive jurisdiction.
Jonas Weber
Jonas Weber
CEO · Nimbus Cloud GmbH
Date: 02.09.2026
M. Fernández
Marta Fernández
COO · Acme Operations GmbH
Date: 02.09.2026
Nimbus × Acme · ConfidentialPage 2 of 2
fx=SUM(E4:E8)
ABCD
1
Payroll · September 2026
2 
3456789
EmployeeHoursGrossNet
J. Carter1688,333.336,212.40
L. García1607,100.005,402.75
A. Okafor1726,880.005,236.10
M. Fernández1505,400.004,180.20
S. Tanaka1647,950.005,970.35
Total81435,663.3327,001.80
Approved by Finance · 30 Sep 2026. Net pay is after income tax and social security. Transfers scheduled for 1 Oct 2026.
SeptemberAugustSummary
NNimbus Cloud GmbHCloud infrastructure
INVOICE
From
Nimbus Cloud GmbH
42 Speicherstrasse · 10115 Berlin
Bill to
Acme Operations GmbH
8 Marktgasse · 1010 Vienna
Invoice no.INV-2026-0718
Issued30 July 2026
Due13 August 2026
DescriptionQtyAmount
Cloud hosting subscriptionJuly 2026 · Production1€1.200,00
Subtotal€1.200,00
VAT 20%€240,00
Total due€1.440,00
Payment details
Bank transfer · IBAN DE89 3704 0044 0532 0130 00
Reference INV-2026-0718
Nimbus Cloud GmbH · Fictional demonstration invoicePage 1 of 1
CALIFORNIA COMMERCIAL LEASE AGREEMENT
This Lease Agreement made the 2025-05-20, by and between Hope Mante [name of lessor], of 925 The Mews [street address], hereinafter referred to as "Lessor", and Reilly Group [name of lessee], hereinafter referred to as "Lessee".
1. DESCRIPTION OF LEASED PREMISES
The Lessor agrees to lease to the Lessee 2448 square feet of Industrial Space located at 80369 Efren Burg, State of California.
2. TERM OF LEASE
The term of this Lease shall be 4 year(s) 5 month(s), commencing on the 20 May 2025 and expiring at midnight on 20 October 2029.
3. BASE RENT
The net monthly payment shall be 8646 dollars per month. A security deposit of 1395 dollars is required.
☑ Lessee may not renew the Lease.
☐ Lessee may renew with written notice 60 days prior.
Page 5 of 8
REPUBLIC OF ALDORIAIDENTITY CARD
Surname
MANTE
Given names
HOPE ELISE
Date of birth
14.03.1988
Sex
F
Document no.
AX4821937
Date of expiry
22.09.2031
H. Mante
Address
925 THE MEWS, ALDOR CITY
Issuing authority
CITY OF ALDOR · 22.09.2021
Height · Eyes
168 CM · BROWN
IDALDAX48219376<<<<<<<<<<<<
8803145F3109225ALD<<<<<<<<4
MANTE<<HOPE<ELISE<<<<<<<<
{
"invoice_no": "INV-2026-0718","due_date": "2026-08-13","total_due": 1440.00}

Parse, Extract, Lineage

Parse

POST /parse

PDF or imageelements + markdown

open-weight hosted, from
$1/ 1,000 pages
Compare OCR prices

Extract

POST /extract

parse + JSON schematyped JSON

no per-document fee
Per tokenat your chosen model’s rate
Extraction API

Lineage

grounding: "field"

extracted fieldsource region + history

with every grounded extract
$0included
Explore lineage

Supported OCR providers

  • PaddleOCR-VL
  • Mistral OCR
  • Azure Document Intelligence
  • Google Document AI
  • Amazon Textract

Check the source of a value

Inspect source text, model output, and edits in the lineage graph.

commercial-lease.jpg
1 page · 3 fields
California commercial lease agreement with handwritten entries
Data residency

EU-only processing

Contact us to arrange EU processing for your document volume.

Contact usHow it works

EU-only processing

Keep documents and results in EU regions.

EU-hosted models

We run open-weight models on EU GPUs and use cloud providers' EU regions.

No data retention

We delete files and results after processing and do not use them for training.

Compare 9 OCR models

Choose an ocr_model for POST /parse. Prices below are per 1,000 pages.

  • PaddleOCR-VL 1.6paddleocr-vl-1.6

    Open-weight hosted. General documents at the lowest page price.

    ProviderBaidu / Paddle
    Pricing / 1,000 pages
    $1
    OutputMarkdown + boxes
  • Mistral OCR 4mistral-ocr-4

    Complex, visually rich docs when quality matters more than price.

    ProviderMistral
    Pricing / 1,000 pages
    $4.60
    OutputMarkdown + boxes
  • Mistral OCR 3mistral-ocr-3

    Clean markdown without bounding-box regions.

    ProviderMistral
    Pricing / 1,000 pages
    $2.30
    OutputMarkdown
  • Azure DI Layoutazure-di-layout

    Microsoft-centric deployments and layout-heavy enterprise PDFs.

    ProviderMicrosoft Azure
    Pricing / 1,000 pages
    $11.50
    OutputMarkdown + boxes
  • Azure DI Readazure-di-read

    Search and indexing in Microsoft environments.

    ProviderMicrosoft Azure
    Pricing / 1,000 pages
    $1.73
    OutputPlain text + boxes
  • Google Enterprise Document OCRgoogle-docai-ocr

    Google Cloud deployments needing structured OCR.

    ProviderGoogle Cloud
    Pricing / 1,000 pages
    $1.73
    OutputMarkdown + boxes
  • AWS Textract Detectaws-textract-detect

    Low-cost text detection for AWS-native pipelines.

    ProviderAWS
    Pricing / 1,000 pages
    $1.73
    OutputPlain text + boxes
  • AWS Textract Layoutaws-textract-layout

    Headings, sections, and layout-aware reading order on AWS.

    ProviderAWS
    Pricing / 1,000 pages
    $4.60
    OutputMarkdown + boxes
  • AWS Textract Tables + Layoutaws-textract-tables-layout

    Forms and financial docs with important tables in AWS.

    ProviderAWS
    Pricing / 1,000 pages
    $17.25
    OutputMarkdown + tables + boxes

Parse PDFs, then extract fields

Try a request with cURL, TypeScript, or Python.

Install
$ npm i @openparser/sdk
$ pip install openparser-sdk
Full API referenceCreate an API key
POST/parserequest
curl https://api.openparser.dev/parse \
  -H "Authorization: Bearer $OPENPARSER_API_KEY" \
  -H "Idempotency-Key: $(uuidgen)" \
  -F 'request={"ocr_model": "paddleocr-vl-1.6"};type=application/json' \
  -F "file=@invoice.pdf"
200 · resultresponse
{
  "document_id": "doc_3f9a2c",
  "elements": [
    {
      "id": "elem-total",
      "kind": "text",
      "text": "Total due: €1.440,00",
      "locations": [{
        "page_number": 1,
        "bbox": { "left": 468, "top": 712, "right": 1118, "bottom": 760 }
      }],
      "confidence": { "score": 0.98 }
    }
  ],
  "markdown": "# INVOICE\n…"
}

Try it in Studio

Upload a document and inspect the parsed text and extracted fields.

Upload a document, pick an OCR model, and switch on extraction.

Open StudioCreate an API key

You pay per parsed page

Start free, then pay only for the pages you parse and the tokens you extract.

free credits
$10

At signup. About 10,000 pages on open-weight models.

pay per page
$1/ 1,000 pages

On open-weight models. Cloud OCR bills at its own page rate.

extraction
Per token

Billed on the tokens of the llm_model you pick.

  • $10 free credits to start
  • No minimums
  • No per-seat fees
  • Sync, async, and batch
  • Source citations
  • Playground and API keys
Create an API key

Working with the API

Choose your models and use a shared document format.

Choose an OCR model

Set ocr_model per request, including sync, async, and batch calls.

Pay per page

Hosted open-weight models start at $1 per 1,000 pages.

OpenParser document output

Read the same element fields across OCR models.

Separate extraction LLM

Set llm_model to fill your JSON schema from a parsed document.

Citations to source elements

Use element_id citations to highlight the source region on a page.

Open schema and adapters

Use @openparser/schema and @openparser/adapters in hosted or self-hosted projects.

Frequently asked questions

What is OpenParser?

You send a PDF to the REST API, choose among nine hosted OCR engines, and receive the same OpenParser elements and markdown. You pay per page.

Which OCR models can I use?

Nine public models today: paddleocr-vl-1.6, mistral-ocr-4, mistral-ocr-3, azure-di-layout, azure-di-read, google-docai-ocr, aws-textract-detect, aws-textract-layout, and aws-textract-tables-layout. Discover live pricing and options with GET /models/ocr.

How much does OpenParser cost?

OCR bills per page: open-weight hosted models start at $1 per 1,000 pages (for example PaddleOCR-VL), and other models list their own rates on the pricing page. Extract bills the tokens of the LLM you choose. Lineage is included with every grounded extract at no extra cost. New accounts get $10 in free credits.

What does the document parsing API return?

Every model returns the same OpenParser document graph: provenance, pages, elements, reading-order text, and best-effort markdown. Element locations can include page numbers and bounding boxes. Scoped OCR confidence is included when the engine supplies it.

What is the difference between OCR, Extract, and Lineage?

OCR (POST /parse) turns a document into an OpenParser document graph of elements and markdown. Extract (POST /extract) fills your JSON schema from that graph using the LLM you pick. Lineage comes back when you set grounding to field: each value cites the element it came from, plus the model and review steps that produced it.

Can I process documents asynchronously or in batches?

Yes. The API supports synchronous, asynchronous, and batch parsing and extraction, plus job endpoints for checking status and retrieving results.

Do I change client code when I switch OCR providers?

Usually no. Change ocr_model on the request and keep the same response shape. For self-hosted or direct provider calls, use the open-source @openparser/schema and @openparser/adapters packages.

Can OpenParser process documents in the EU?

The hosted API runs in AWS US regions today. EU-only processing and self-hosted deployments are available on request for teams with GDPR or DORA data-location requirements. Book a call from the EU processing page and tell us your volume and requirements.

$10 free credits

Try your own document

Upload a PDF to Studio to get started.

Open StudioCreate an API key