Ship document processing in minutes
REST APIs for OCR, PII masking, and structured extraction. Get an API key, install an SDK, and make your first call the same afternoon. Synchronous for small documents, asynchronous with webhook callbacks for large jobs.
From zero to first response
Four steps. No infrastructure to provision, no model to train. The free trial gives every organisation a one-time Rs 500 credit with no card required.
A request looks like this
Authenticate with your API key in a header and send the file. The example below posts a scanned PDF to the OCR Full-Text endpoint.
curl -X POST https://api.abscode.com/v1/ocr \
-H "Authorization: Bearer $ABSCODE_API_KEY" \
-F "file=@invoice.pdf" \
-F "searchable_pdf=true"
# Response (synchronous, small document)
{
"job_id": "ocr_8f3c2a",
"status": "completed",
"pages": 2,
"text": "TAX INVOICE ...",
"confidence": 0.97,
"blocks": [
{ "page": 1, "text": "TAX INVOICE", "bbox": [62, 40, 318, 78] }
]
}Illustrative — see live docs for exact schema
Field names, endpoint paths, and parameters above are for orientation only. Always refer to the live reference for the authoritative request and response schema before integrating.
Use your language
Official SDKs handle authentication, retries, and the asynchronous polling and webhook flow so you write less plumbing.
Python
pip install and call an endpoint in a few lines. Async helpers wait on webhook or polling for you.
Node
Install via npm. Promise-based client with typed helpers for OCR, masking, and extraction.
Java
Maven and Gradle artifacts for JVM services. Synchronous and async request builders included.
.NET
NuGet package targeting modern .NET. Strongly typed response models for extracted fields.
PHP
Composer package for PHP applications. Handles multipart upload and webhook verification.
cURL
No SDK required. Every endpoint is plain HTTPS, so you can integrate from any language or shell.
What the APIs return
Five endpoints, one consistent response model. This overview orients you — the live reference holds the authoritative schema.
Text and searchable PDF
Full text extraction with auto-deskew and orientation correction. Returns plain text plus an optional searchable PDF. Multi-language including Hindi, Tamil, Telugu, and Arabic. Rs 0.25 /page.
OCR Full-Text →Redaction in place
Masks Aadhaar, PAN, account numbers, names, addresses, phone, and signatures with configurable rules. Output preserves the original layout. Rs 0.25 /page.
PII Masking →Structured fields as JSON
Named fields with bounding boxes from invoices, statements, receipts, KYC forms, and contracts. Basic at Rs 0.90 /page; Pro adds confidence scores, guardrails, and validation rules at Rs 1.40 /page.
Extract Pro →Rule-based review
Custom, project-based analysis: NDA gap analysis, contract compliance, policy adherence, and regulatory filing validation. Configured to your documents and routed through sales.
Document Analysis →Sync vs async
Small documents return synchronously in the same response. Larger or multi-page jobs run asynchronously and POST a callback to your webhook URL when complete.
Response shape
Structured JSON across all endpoints: extracted data, confidence values, and bounding boxes that map each field back to its position on the page.
Supported formats
PDF, TIFF, JPG, PNG, and scanned images, single or multi-page. Pre-processing applies deskew, denoise, orientation correction, and contrast normalization.
Service levels
99.5% uptime SLA with p95 latency under 8 seconds for a 10-page document. Documents are encrypted with TLS 1.3 in transit and AES-256 at rest, and are auto-purged after processing — we do not retain documents and do not train models on customer data.