AI document extraction

Documents in.Data out.

DocParse turns PDFs, images, and DOCX files into clean JSON. Define the fields you need, upload the file, get structured data back—in any language, no templates required.

Start extractingBook a demo
100 pages free every month No card required Pay in USD or INR
app.docparse.io / extractions / inv-20294
Live
ACME SUPPLY CO.
Invoice #INV-20294
Bill toNorthwind Logistics
DateApril 14, 2026
DueMay 14, 2026
DescriptionQtyTotal
Steel pallet, 48"12$840.00
Strapping, polyester4$162.00
Pallet wrap, 80g8$320.00
Freight, ground1$520.50
Total due$1,842.50
JSONCSVWebhook98.6% confidence
{
"vendor": "Acme Supply Co.",
"invoice_no": "INV-20294",
"bill_to": "Northwind Logistics",
"date": "2026-04-14",
"line_items": [
    { "Steel pallet", 12, 840.00 },
    { "Strapping", 4, 162.00 }
],
"total": 1842.50,
"currency": "USD"
}
Extracted in 2.4s
Scroll to see how
Watch it work

Fields lift off the page.
Land in your schema.

Tell DocParse which fields you need. It reads the document and returns them as clean JSON—ready for your database, webhook, or Zap.

ACME SUPPLY CO.
Invoice #INV-20294
vendorAcme Supply Co.
invoice noINV-20294
bill toNorthwind Logistics
date2026-04-14
total$1,842.50
extracted.json
{
  "vendor": "Acme Supply Co.",
  "invoice_no": "INV-20294",
  "bill_to": "Northwind Logistics",
  "date": "2026-04-14",
  "total": "$1,842.50"
}
✓ Extracted in 2.4s
Workflow

Define your fields. Get the data.

You decide what to pull out: vendor names, totals, dates, line items, anything. DocParse fills your schema in seconds—no templates, no training, no fine-tuning.

01
Define
Set the fields you want to extract. Type names and descriptions in the dashboard, or push them through the REST API.
02
Upload
Drop PDFs, images, or DOCX files. Up to 25 MB per file, 30 files per batch. Mix layouts and languages in the same batch.
03
Extract
Get clean JSON back. Download as CSV, listen on a signed webhook, or pipe into Zapier.
Intelligence

Powered by frontier
multi-modal AI.

DocParse reads layouts it has never seen before. You don't train it. You don't template it. You give it a document and tell it which fields you want.

"vendor"
"total"
"date"
No templates
Define the fields you need in plain language. DocParse adapts to each document on the first run—no per-vendor setup.
EN日本ES
Any language
Multi-lingual by default. Pick a specific language or leave it on auto—handwriting, mixed scripts, and right-to-left all supported.
Tables & checkboxes
Toggle the table, checkbox, chart, or handwritten flags and DocParse returns rich nested objects, not flat strings.
2.4s
Extract + Classify
Two workflows in one product. Pull structured fields, or sort incoming documents into categories you define—from the same API.
Coverage

Any document.
Any language. Any layout.

PDFs, JPGs, PNGs, WEBP, and DOCX—up to 25 MB per file. Scans, photos, and digitally generated documents all go through the same pipeline.

Invoices
Vendor, line items, totals, taxes.
Receipts
Merchant, items, totals, payment method.
Contracts
Parties, dates, clauses, signatures.
P<USA<DOE<<JANE
123456789USA
Resumes
Skills, experience, education, contact.
Bank checks
Payee, amount, routing, signatures.
Business cards
Name, title, company, contact details.
Emails
Sender, subject, body, parsed attachments.
Custom fields
Define your own schema. DocParse fills it.
Built for developers

One REST API.
Any document.

Plain HTTP. Bearer-token auth with revocable keys. Signed outbound webhooks for async pipelines, and a native Zapier app so non-developers can ship too.

Read the docs Get an API key
Works withRESTWebhooksZapier
# 1. Create an extraction with the fields you want
curl https://api.docparse.io/api/v1/createExtraction \
  -H "Authorization: Bearer $DOCPARSE_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Invoices",
    "template": "invoice",
    "fields": [
      { "name": "vendor", "type": "string" },
      { "name": "total", "type": "number" },
      { "name": "due_date", "type": "date" }
    ]
  }'

# 2. Upload a file to it
curl https://api.docparse.io/api/v1/uploadFiles \
  -H "Authorization: Bearer $DOCPARSE_KEY" \
  -F "extractionId=ext_..." \
  -F "files=@invoice.pdf"
Trust

Security by design.

Your documents stay yours. TLS everywhere, hashed API keys, signed webhooks, and a processing pipeline that never retains your data for model training.

Never trained on your data
Your documents go to the model, the structured output comes back, and that's it. Nothing is retained for training — ours or any third party's.
vendor•••••
total••••
iban•••••••
Encrypted in transit
TLS on every request to the API, dashboard, and webhook endpoints. File downloads use short-lived signed URLs that expire after one hour.
HMACSHA-256
Signed webhooks
Every outbound webhook is signed with HMAC-SHA256 and a per-endpoint secret, following the Standard Webhooks spec. Replay-protected by timestamp.
dp_•••
Hashed, revocable keys
API keys are stored as SHA-256 hashes — we can't see them after creation. Rotate or revoke instantly from the dashboard, with usage logs per key.
FAQ

Questions, answered.

Do I need to set up a template first?
You define the fields you want — name, type, optional description — and DocParse handles the rest. There are built-in templates for invoices, receipts, contracts, resumes, bank checks, business cards, bills of lading, and emails, plus a free-form custom template for anything else.
Which file types are supported?
PDF, PNG, JPG / JPEG, WEBP, DOCX, and plain text. Up to 25 MB per file and 30 files per batch. Mix layouts and languages freely in the same batch.
Which languages work?
Any language the underlying multi-modal model supports. The default is multi-lingual so a mixed-script document works without extra config — you can also pin a specific language per extraction. Handwriting and right-to-left scripts are supported.
How does pricing work?
Every account gets 100 pages every month free, forever. Beyond that, pick a pay-as-you-go package (one-time top-up, pages don't expire) or a monthly subscription at a lower per-page rate. Nine volume tiers from 100 to 50,000 pages, in USD or INR.
How do I get the data back?
Three options: view and export from the dashboard as JSON or CSV; poll the REST API for finished batches; or register a webhook endpoint and we'll push signed deliveries (HMAC-SHA256) the moment an extraction finishes.
Can I use this without writing code?
Yes. The dashboard handles uploads, field definitions, results, and exports end-to-end. For workflows, our Zapier app lets you trigger downstream actions in 6,000+ apps when an extraction completes — no code required.
Ready when you are

Start extracting
in under a minute.

100 pages free, every month. No credit card. Pay in USD or INR when you need more.

Try it freeBook a demo