​

Income Tax OCR APIs

​

Supported Documents

​

How It Works

1. Upload Document: Submit PDF or image file
2. Processing: System extracts and classifies data
3. Get Results: Receive structured JSON with confidence scores

​

Related Documentation

• Calculator APIs
• Report APIs

{
  "document_type": "form_16",  
  "file": "Base64 encoded PDF/Image",
  "taxpayer_pan": "AAAPI0000A",
  "financial_year": "2024-25",
  "confidence_threshold": 0.85
}

​

Processing Flow

1. Upload Document: Submit PDF/image with metadata
2. Initial Classification: Auto-detect document type and version
3. OCR Processing: Extract text and tabular data using advanced AI models
4. Data Validation: Validate extracted fields against expected patterns
5. Normalization: Standardize dates, numbers, and formats
6. Return Results: Return structured JSON with confidence scores

​

Response Structure

​

Successful Extraction

{
  "code": 200,
  "data": {
    "document_type": "form_16",
    "confidence": 0.95,
    "extraction_details": {
      "employer": {
        "name": "ABC Technologies Pvt Ltd",
        "pan": "AABCT1234K",
        "address": "Bangalore, Karnataka"
      },
      "employee": {
        "name": "Rajesh Kumar",
        "pan": "BXRPK5678A",
        "designation": "Senior Software Engineer"
      },
      "salary": {
        "gross_salary": 2400000,
        "basic": 1000000,
        "hra": 600000,
        "dearness_allowance": 200000,
        "other_allowances": 400000,
        "deductions": 100000,
        "net_salary": 2300000
      },
      "tax_details": {
        "tds_deducted": 350000,
        "surcharge": 15000,
        "cess": 2000,
        "total_tax": 367000
      },
      "investment_declarations": [
        {
          "section": "80C",
          "investment_type": "Life Insurance",
          "amount": 150000
        }
      ],
      "fy": "2024-25",
      "issue_date": "2025-03-31"
    },
    "extraction_metadata": {
      "pages_processed": 1,
      "extraction_timestamp": "2025-01-20T10:30:00Z",
      "model_version": "v2.1"
    }
  },
  "timestamp": 1750687659809,
  "transaction_id": "txn_ext_123456"
}

​

Confidence Scores

{
  "employer_name": {
    "value": "ABC Technologies Pvt Ltd",
    "confidence": 0.98
  },
  "tds_deducted": {
    "value": 350000,
    "confidence": 0.92
  }
}

• 0.95+: Very high confidence, can be used directly
• 0.85-0.95: High confidence, minimal validation needed
• 0.75-0.85: Moderate confidence, recommend review
• <0.75: Low confidence, manual verification recommended

​

Partial Extraction

{
  "code": 200,
  "data": {
    "document_type": "form_16",
    "confidence": 0.78,
    "extraction_details": { },
    "warnings": [
      {
        "field": "investment_declarations",
        "reason": "Page quality too low",
        "recommendation": "Manual verification suggested"
      }
    ]
  }
}

​

Integration Patterns

​

Customer KYC/Income Verification

1. Accept Form 16 upload from customer
2. Extract income and employer details
3. Verify against other sources
4. Use for lending decisions or account opening

​

Tax Compliance Workflows

1. Employee uploads Form 16 from multiple employers
2. System extracts TDS details from each
3. Verify total TDS matches Form 26AS (reconcile discrepancies)
4. Auto-populate ITR with extracted data
5. Flag any mismatches for manual review

​

Payroll Integration

1. HR uploads employee Form 16 PDFs in bulk
2. Extract salary information and TDS
3. Validate against payroll records
4. Reconcile any discrepancies
5. Generate compliance reports

​

Document Archival

1. Upload historical tax documents
2. Extract and index searchable data
3. Create audit trail of document processing
4. Enable quick retrieval and verification

​

Async Processing Pattern

1. Submit Document: POST upload request → Get job_id
2. Check Status: GET job status endpoint → Receive processing progress
3. Fetch Results: GET results endpoint → Download extraction results

• Processing of large documents without timeout
• Bulk processing of multiple documents
• Progress tracking for UI feedback
• Retry mechanisms for failed extractions

​

Error Handling & Validation

​

Common Issues & Solutions

​

Quality Checks

• Page Count: Verify all pages uploaded (multi-page documents)
• Document Completeness: Ensure all required fields present
• Data Validity: Validate extracted dates, amounts, PAN formats
• Cross-field Validation: Verify relationships between fields (e.g., salary components sum to gross)

​

Best Practices

• Document Quality: Upload clear, well-lit scans or digital PDFs for best results
• Complete Documents: Ensure all pages included, especially if multi-page
• Metadata: Provide taxpayer PAN and financial year for better validation
• Confidence Thresholds: Set appropriate confidence thresholds for your use case
• Validation: Always validate extracted data, especially for high-value decisions
• Feedback: Report extraction errors to improve model accuracy over time
• Privacy: Handle sensitive documents securely, follow data privacy regulations

​

Related Documentation

• Form 16 Extraction API
• Form 26AS Extraction API
• Report Generation from Extracted Data
• IT Context Guide