Extract structured data from PDFs, invoices, receipts, and forms with a single API call. Send any document, get clean JSON back. No templates, no training data, no configuration files.
Send a document to the extraction endpoint and receive structured JSON. Define a schema to control exactly which fields you need, or let the API auto-detect the document structure.
curl -X POST \
https://api-parse.conversiontools.io/v1/extract \
-H "Authorization: Bearer YOUR_API_KEY" \
-F "file=@invoice.pdf" \
-F 'schema={
"vendor": "string",
"date": "string",
"total": "number",
"line_items": [{
"description": "string",
"amount": "number"
}]
}'{
"status": "completed",
"data": {
"vendor": "Acme Corp",
"date": "2026-03-01",
"total": 1250.00,
"line_items": [
{
"description": "Consulting services",
"amount": 1000.00
},
{
"description": "Expenses",
"amount": 250.00
}
]
}
}Three steps to go from raw document to structured data. No training, no templates - the API understands your documents automatically.
Send any PDF, image, or scanned document to the API endpoint. Supports invoices, receipts, forms, contracts, and more.
The API reads the document, understands its structure, and extracts the fields you requested in your schema definition.
Receive clean, typed JSON matching your schema. Ready to store in a database, feed into a pipeline, or display in your application.
Built for developers who need reliable, accurate data extraction without the complexity of traditional OCR pipelines.
Uses large language models to understand document context and layout. Handles varied formats without custom rules or template configuration.
One endpoint, one API call. Send a file, get JSON back. No SDKs required - works with curl, Python, Node.js, or any HTTP client.
Define exactly which fields to extract with JSON schemas. Support for strings, numbers, dates, arrays, and nested objects. Reuse schemas across documents.
Process PDFs, scanned images, JPEGs, PNGs, GIF, WebP, TIFF, BMP, HEIC, and AVIF files. Works with both digital and scanned documents across all languages.
Files are processed and deleted automatically. No document data is stored after extraction. EU-hosted infrastructure with encrypted connections.
Most single-page documents are processed in seconds. Synchronous and asynchronous modes available depending on your use case and document size.
Most documents are not clean, single-column text. The API reads the messy ones too - tables, scans, and pages that run long.
Recognizes tabular data, bulleted lists, numbered clauses, and nested structures, and returns each row or item as a typed object in a JSON array.
Extracts printed, typed, and handwritten content from scanned documents, including annotations and filled-in form fields.
Processes an entire document in one call and keeps context across pages, so values that span headers, sections, and appendices come back together.
You Got the Data Out. Why Is It Still a Mess?
Parsing a document is the easy part now. The real work is turning that output into the same clean fields every time - here is the gap that breaks pipelines, and how to close it.
Read moreHow I prepare my VAT report from a pile of supplier invoices
A founder walkthrough: define the fields once, then turn every supplier PDF invoice into the same structured JSON for a VAT return - no manual re-typing.
Read moreThe same schema-driven API works across every document type. Define a schema once, extract from thousands of files.
Parse any PDF into structured JSON, scanned or digital.
Vendor, line items, totals, tax, and dates from invoices.
Store, items, totals, and payment method from receipts.
PO number, vendor, buyer, and SKU-level line items.
Transactions, running balances, and dates for reconciliation.
Box-level data from W-2, 1099, and other tax forms.
Parties, dates, governing law, and key clauses.
Carrier, parties, ports, containers, and cargo.
Common questions about the data extraction API.
Parse accepts PDF, JPEG, PNG, GIF, WebP, TIFF, BMP, HEIC, and AVIF files. You can extract structured data from scanned documents, digital PDFs, photos of receipts, and any image-based document.
No. Parse uses AI to understand document structure automatically. You define a JSON schema describing the fields you want, and the API extracts them from any document - no templates, training data, or configuration files required.
Parse uses large language models to understand document context, not just OCR text matching. This means it handles varied layouts, languages, and formatting. Accuracy depends on document quality, but most structured documents like invoices and receipts achieve high extraction rates.
Yes. The free plan includes 100 pages per month with full API access, custom schemas, and all supported file formats. No credit card required to start.
For batch processing, use the asynchronous extraction endpoint. Submit documents and poll for results, or use webhooks to get notified when extraction is complete. The Pro plan supports 5,000 pages per month with priority processing.
Both. Use the no-code dashboard to define your fields and preview the parsed results in your browser, or call the REST API to automate the same extraction at scale. The parsing logic is identical - the dashboard is just a friendly front end over the API.
No SDK to install. Parse is a plain REST API, so you can call it from any language - curl, Python, Node.js, Go, or any HTTP client. Send a file and your JSON schema to a single endpoint and get structured data back.
Get your API key and extract data from your first document in minutes. 100 pages per month free - no credit card required.