BharatParse API

Overview

BharatParse is a document extraction API built specifically for Indian businesses. Send any invoice, bill, or receipt as a base64-encoded file and receive clean, validated JSON in return - with confidence scores, field-level warnings, and GST-specific validation built in.

Why BharatParse vs AWS Textract / Google Vision

Quick Start

Get your first extraction in under 2 minutes.

1. Get your API key

Subscribe to any plan on RapidAPI. Your API key will be shown in the dashboard.

2. Make your first request

# Python example
import base64, requests

# Load your document
with open("invoice.pdf", "rb") as f:
    b64 = base64.b64encode(f.read()).decode()

# Call the API
response = requests.post(
    "https://bharatparse-indian-invoice-bill.p.rapidapi.com/v1/extract",
    headers={
        "X-RapidAPI-Key": "YOUR_RAPIDAPI_KEY",
        "X-RapidAPI-Host": "bharatparse-indian-invoice-bill.p.rapidapi.com",
        "Content-Type": "application/json"
    },
    json={
        "file_b64": b64,
        "file_type": "pdf",
        "schema": "auto",
        "country": "IN"
    }
)
print(response.json())

3. That's it

You'll receive structured JSON with all extracted fields, a confidence score, and any warnings about low-quality fields.

Authentication

BharatParse uses RapidAPI's standard authentication. Include your RapidAPI key in every request header:

Supported File Formats

BharatParse accepts both PDFs and images. You can send a scanned document, a phone photograph, a screenshot, or a downloaded PDF - all are handled by the same endpoint.

File size limit

Maximum 20MB per request. Most invoices and bills are well under 1MB. Bank statements and multi-page PDFs are typically 2–5MB. If your file exceeds 20MB, compress the PDF or reduce image resolution before sending.

Endpoint

Request Parameters

All parameters are sent as a JSON body.

Response Format

All successful responses (HTTP 200) follow this structure:

{
  "schema_detected": "restaurant",     // Document type identified
  "confidence": 0.92,               // 0.0–1.0 extraction quality score
  "data": { ... },                   // Extracted fields (schema-specific)
  "warnings": [                      // Field-level issues (empty if none)
    "Invoice date not visible in scan"
  ],
  "processing_ms": 1843            // Processing time in milliseconds
}

Confidence Score Guide

Document Schemas

Use the schema parameter to specify the document type, or use auto to let BharatParse detect it automatically.

Example: Restaurant Bill

Input: a blurry, angled phone scan of a Starbucks receipt. The OCR text was heavily degraded - yet all key fields extracted correctly.

Input document

Restaurant_020525.pdf - page 1

Extracted JSON

{ "schema_detected": "restaurant", "confidence": 0.92, "data": { "restaurant_name": "Starbucks", "hsn_code": "996331", "line_items": [ { "name": "Tall Cold Coffee", "quantity": 1, "total": 320.00 } ], "taxable_value": 320.00, "cgst_rate": 2.5, "cgst_amount": 8.00, "sgst_rate": 2.5, "sgst_amount": 8.00, "grand_total": 336.00, "payment": { "mode": "starbucks_card", "card_last4": "1821" } }, "warnings": ["Invoice date not visible in scan"], "processing_ms": 1843 }

Example: Handwritten Fuel Receipt

Input: a handwritten BPCL cash memo where the quantity "94:14" was written with a colon instead of a decimal point. BharatParse interpreted the context and extracted the correct value.

Input document

050625fuel.pdf - handwritten pump memo

Extracted JSON

# Response for handwritten BPCL receipt
{
  "schema_detected": "fuel",
  "confidence": 0.90,
  "data": {
    "dealer_name": "N. M. Shamsuddin & Sons",
    "oil_company": "BPCL",
    "invoice_date": "2025-06-05",
    "fuel_items": [
      {
        "fuel_type": "Speed",
        "litres": 94.14,
        "rate_per_litre": 21.24,
        "amount": 2000.00
      }
    ],
    "total_amount": 2000.00
  },
  "warnings": ["Litres value '94:14' is handwritten and interpreted as 94.14"],
  "processing_ms": 8598
}

Example: Telecom Bill

Input: a 7-page Jio Fiber PDF bill. BharatParse extracted only the billing summary - ignoring 80 rows of itemised data usage - and returned the fields that actually matter for accounting.

Input document

SEPTEMBER2025.pdf - Jio Fiber bill (page 1 of 7)

Extracted JSON

# Response for Jio Fiber bill
{
  "schema_detected": "telecom",
  "confidence": 1.0,
  "data": {
    "provider": "Jio",
    "customer_name": "Mr. Rajesh Kumar",
    "account_number": "XXXXXXXXXXXX9305",
    "due_date": "2025-09-30",
    "plan_name": "Postpaid_399_6M: Unlimited Data @ 30 Mbps",
    "vendor_gstin": "24AABCI6363G1ZP",
    "charges": {
      "current_taxable_charges": 399.00,
      "cgst_rate": 9.0, "cgst_amount": 35.91,
      "sgst_rate": 9.0, "sgst_amount": 35.91
    },
    "total_payable": 470.82
  },
  "warnings": [],
  "processing_ms": 24406
}

Example: IRCTC Train Ticket

Input: an IRCTC ERS PDF with the GST invoice on page 2. BharatParse extracted passenger details from page 1 and the full GST breakdown from page 2 in a single API call.

Input document

24th_Jan.pdf - IRCTC ERS ticket

Extracted JSON

# Response for IRCTC Tejas Express ticket
{
  "schema_detected": "travel",
  "confidence": 0.95,
  "data": {
    "pnr": "8543381796",
    "train_number": "82902",
    "train_name": "IRCTC TEJAS EXP",
    "journey_date": "2026-01-24",
    "from_station": "AHMEDABAD JN (ADI)",
    "boarding_station": "VADODARA JN (BRC)",
    "to_station": "BORIVALI (BVI)",
    "passengers": [
      { "name": "RAJESH KUMAR", "age": 67, "current_status": "WL/44" }
    ],
    "fare": { "ticket_fare": 1680.00, "total_fare": 1715.40 },
    "gst": { "igst_rate": 5.0, "igst_amount": 80.00, "total_tax": 80.00 }
  },
  "warnings": [],
  "processing_ms": 11990
}

Example: GST Invoice

Full B2B tax invoice extraction with automatic GSTIN checksum validation.

# Response for B2B GST invoice
{
  "schema_detected": "gst_invoice",
  "confidence": 0.97,
  "data": {
    "invoice_number": "INV-2024-009182",
    "invoice_date": "2024-11-15",
    "vendor": {
      "name": "Tata Power Ltd",
      "gstin": "27AAACT2727Q1ZW",
      "pan": "AAACT2727Q"
    },
    "line_items": [
      {
        "description": "IT Services",
        "hsn_sac": "998313",
        "taxable_amount": 50000.00,
        "cgst_rate": 9, "cgst_amount": 4500.00,
        "sgst_rate": 9, "sgst_amount": 4500.00
      }
    ],
    "totals": { "taxable_value": 50000, "grand_total": 59000 }
  },
  "warnings": [],
  "processing_ms": 2100
}

Error Codes

Error Response Format

{
  "detail": "Unsupported file_type 'bmp'. Use: pdf, jpeg, jpg, png, webp, tiff, tif"
}

Plans & Pricing

Subscribe at rapidapi.com. All plans include all 13 document schemas, confidence scores, and GST validation.

FAQ

What file formats are supported?

BharatParse accepts both PDFs and images. Supported formats: PDF, JPEG (.jpeg or .jpg), PNG, WebP, and TIFF (.tiff or .tif). Maximum file size is 20MB per request. This means you can send a phone photo of a receipt (JPEG), a scanned document (PDF or TIFF), or a screenshot (PNG) - all work equally well. For best accuracy, use PDF or high-resolution JPEG at 150 DPI or above.

Does it work with handwritten receipts?

Yes. BharatParse uses an LLM rather than traditional OCR. This means it reads documents the way a person would - understanding context rather than just recognising characters. The fuel schema is specifically tuned for handwritten pump memos. Confidence scores will reflect extraction uncertainty.

How accurate is GSTIN validation?

Every GSTIN goes through full 15-character format validation and checksum verification automatically. Invalid GSTINs are flagged in the warnings array rather than silently passed through.

What happens when the document type is unclear?

Use schema: "auto". The model will identify the document type and apply the appropriate extraction schema. The detected type is returned in schema_detected.

Can I use it for non-Indian documents?

Yes - set the country parameter to your ISO country code (e.g., AE for UAE, SG for Singapore). The extraction adapts tax field names for that country, though India-specific checks like GSTIN validation won't apply.

Is my data stored?

No. Documents are processed in memory and discarded immediately after extraction. Nothing is logged or stored.

What's the average response time?

Typically 2–10 seconds depending on document complexity and size. Simple single-page bills average around 2 seconds. Multi-page PDFs like bank statements take longer.