DOCUMENT AI · FOR EVERY TEAM

Document technology, absolute for all.

One hosted API for OCR, data extraction, document analysis and masking, returning structured JSON you can wire straight into production. Pay-as-you-go pricing and lean processing that moves and stores less data, so startups, integrators and enterprises all run on the same foundation.

Structured JSON Pay-as-you-go Sync + async Zero data retention
CODE EXAMPLE

Extract an invoice in 6 lines

python
node.js
curl
from abscode import DocumentAI

client = DocumentAI(api_key="abs_sk_...")

result = client.extraction.pro(
    file="invoice.pdf",
    document_type="invoice"
)

print(result.fields)
response.json
{
  "invoice_number": "INV-2026-001",
  "date": "2026-06-05",
  "vendor": "Acme Corp Pvt Ltd",
  "vendor_gstin": "27AAAAA0000A1Z5",
  "total_amount": 12450.00,
  "tax_amount": 1900.00,
  "currency": "INR",
  "line_items": [...],
  "confidence": 0.97
}
EVERY API GETS

The same production-grade foundation

Input formats

PDF · TIFF · JPG · PNG · scanned images. Multi-page documents supported natively.

Sync + Async modes

Sync for small docs (instant response). Async with webhook callback for batch / large documents.

Pre-processing built in

Deskew, denoise, orientation correction, contrast normalization. You upload raw, we handle quality.

Multi-language

Documents in any major language. Multilingual content within the same document handled natively.

Privacy by design

Transmitted data scrambled and anonymized. Zero document retention, auto-purged after processing.

Configurable schemas

Add new document types via simple configuration. No model retraining. Static AI models, predictable, consistent.

PRICING

Pay only for what you process

Per-page pricing. Top up your wallet, credit doesn't expire while account is active. Bonus on bulk top-ups.

Affordable to all. Our optimizations cut processing time and the amount of data moved and stored first, fewer server cycles, so the same production accuracy costs less and carries a lighter footprint. That value passes straight through to per-page pricing, whether you send a hundred pages a month or a hundred thousand.

See full pricing →
FAQ

Common questions

What can the Document AI API do?
It is a single hosted REST API that reads documents and returns structured JSON, delivering four capabilities you can use in any combination: OCR full-text extraction, structured data extraction, rule-based document analysis, and data masking. You call it over HTTPS and wire the JSON response straight into production.
Which file formats and document types does the API accept?
It accepts PDF, TIFF, JPG, PNG and scanned images, and processes multi-page documents natively. Structured extraction is tuned for common business documents including invoices, bank statements, receipts, KYC forms, contracts and payslips, and new document types can be added through configuration rather than model retraining.
Can the API OCR scanned documents and return searchable, multi-language text?
Yes. The OCR capability extracts text from PDFs, scans and images and can generate a searchable PDF from them. It handles major languages, including multilingual content within the same document, and applies built-in pre-processing such as deskew, denoise and orientation correction before reading.
Does the API return structured JSON, and does it support asynchronous processing?
Yes. Every capability returns structured JSON with named fields ready for production. A synchronous mode gives an instant response for small documents, while an asynchronous mode with a webhook callback handles batch jobs and large documents.
What is the difference between Basic and Pro extraction?
Basic returns structured JSON with named fields at production-ready accuracy from documents such as invoices, receipts and KYC forms. Pro adds per-field confidence scores, validation rules and guardrails for higher accuracy on complex documents, and is built for fintech, banking, insurance and other regulated workflows.
Can the API automatically mask Aadhaar numbers or other sensitive data?
Yes. The masking capability performs UIDAI-compliant automatic Aadhaar number redaction and preserves the original document layout in the masked output. It runs through the same hosted API as OCR, extraction and analysis.
Does the API store or retain my documents after processing?
No. Documents are auto-purged after processing with zero retention, and data in transit is scrambled and anonymized. The service is designed to move and store less data while still returning the full structured JSON result.
How is the Document AI API priced?
Pricing is pay-as-you-go and charged per page, so you pay only for the pages you process. You top up a wallet whose credit does not expire while the account is active, bulk top-ups earn bonus credit, and free trial credit lets you start without a credit card — see the full breakdown on /pricing/.
THE FULL TOOLKIT

Explore the rest of the Document AI API

One license, every capability — take this and you also get all of these.

Try any API with free trial credit

No credit card. API key in your inbox in minutes.