EU-hosted, end to end
Storage, OCR and AI models run on our own infrastructure in the EU. No third-party model APIs.
Send a document, get structured data back. Built for developers. Ready in minutes, not months.
curl https://api.parseway.dev/extract \
-H "Authorization: Bearer $PARSEWAY_API_KEY" \
-F file=@invoice_0419.pdf \
-F 'schema={
"invoice_number": "string",
"total_due": "number",
"due_date": "date"
}'{
"invoice_number": "INV-2026-0419",
"total_due": 4450.60,
"due_date": "2026-10-31"
}Turn documents into structured data, then derive decisions or classify the data.
A clean API and a web platform to explore your data and manage API keys.
EU-hosted, end to end. Your documents are never used for training and stay yours until you delete them.
No templates to train, no layouts to configure. Describe what you need and integrate in minutes, not sprints.
total_due extracted as 4450.60Tell us what you need in plain JSON. You get exactly those fields back, named the way your code expects them.
Invoices, receipts, contracts, scans. OCR and AI turn them into structured data you can query.
Run decision models on the extracted data to route, approve or flag documents, without writing the rules yourself.
No settings to find, no plan to upgrade to. This is how every request is handled.
Storage, OCR and AI models run on our own infrastructure in the EU. No third-party model APIs.
Your documents and extracted data stay available until you delete them. Delete anytime.
We process your documents to serve your requests. Nothing else.
Parsing a page is the heavy lifting. Decisions on extracted data are cheap, so they're metered separately.
€4 per 1,000 pages
All prices include VAT. A page is one page of a parsed document. A decision is one decision-model call. Unused quota doesn't roll over.