OCR and Mapping Service Pricing Guide
Important: Cloud service pricing changes frequently. Always refer to the official pricing pages linked below for the most current rates. The estimates provided here are for budgeting purposes only and
Official Pricing Resources
AWS Textract (OCR Service)
Primary Resources:
Billing Model:
Per-page pricing (images count as 1 page; PDFs count per page)
Different APIs have different rates (DetectDocumentText vs AnalyzeDocument)
Pricing varies by AWS region
Region Used: Asia Pacific (Mumbai) - ap-south-1
Google Gemini (OCR and Mapping Service)
Primary Resources:
Billing Model:
Token-based pricing (input tokens + output tokens)
Pricing varies by model version (Flash, Pro, etc.)
Free tier available with rate limits
Image/PDF tokenization depends on resolution and content
Models Used:
OCR:
gemini-2.0-flash-exp(configurable)Mapping:
gemini-1.5-flash(configurable)
Amazon Bedrock (Mapping Service)
Primary Resources:
Billing Model:
Token-based pricing (input tokens + output tokens)
Pricing varies significantly by model provider (Anthropic, Meta, Amazon, etc.)
Pricing varies by AWS region
Some models support provisioned throughput pricing
Region Used: Asia Pacific (Mumbai) - ap-south-1
Model Used: meta.llama3-8b-instruct-v1:0 (configurable via OCR_MAPPING_BEDROCK_MODEL_ID)
Service Configuration Details
AWS Textract Configuration
API Used:
DetectDocumentText(synchronous OCR)SDK:
@aws-sdk/client-textractRegion:
ap-south-1(Asia Pacific - Mumbai)Use Case: English language document text extraction
Google Gemini Configuration
API: Google AI Generative Language API (
generativelanguge.googleapis.com)Endpoint:
:generateContentwith inline image/PDF bytesOCR Model: Configurable via
GEMINI_OCR_MODEL(default:gemini-2.0-flash-exp)Mapping Model: Configurable via
OCR_MAPPING_GEMINI_MODEL_NAME(default:gemini-1.5-flash)Use Case: Hindi and mixed-language document processing
Amazon Bedrock Configuration
API:
InvokeModelfor text-to-JSON mappingModel: Configurable via
OCR_MAPPING_BEDROCK_MODEL_ID(default:meta.llama3-8b-instruct-v1:0)Region:
ap-south-1(Asia Pacific - Mumbai)Use Case: Mapping extracted OCR text to structured JSON
Monthly Cost Estimates
Note: These estimates are conservative and intended for budgeting purposes. Actual costs will vary based on document complexity, page count, text length, processing patterns, and current pricing at the time of processing.
Production Assumptions
Average document size: 2 pages per document
All documents processed through selected OCR provider
All documents processed through selected mapping provider
Cost estimates include buffer for retries, failures, and production variability
Exchange rate: 1 USD = 91.46 INR (approximate)
Scenario 1: AWS Textract OCR + AWS Bedrock Mapping
Recommended for: English language documents
Documents per month
AWS Textract OCR USD / INR
AWS Bedrock Mapping USD / INR
Total Monthly Cost USD / INR
1,000
$3 – $8 ₹274 – ₹732
$3 – $15 ₹274 – ₹1,372
$8 – $30 ₹732 – ₹2,744
10,000
$30 – $80 ₹2,744 – ₹7,317
$30 – $150 ₹2,744 – ₹13,719
$60 – $300 ₹5,488 – ₹27,438
100,000
$300 – $800 ₹27,438 – ₹73,168
$300 – $1,500 ₹27,438 – ₹137,190
$600 – $3,000 ₹54,876 – ₹274,380
Pricing References:
Scenario 2: Google Gemini OCR + AWS Bedrock Mapping
Recommended for: Hindi and mixed-language documents
Documents per month
Google Gemini OCR USD / INR
AWS Bedrock Mapping USD / INR
Total Monthly Cost USD / INR
1,000
$2 – $20 ₹183 – ₹1,829
$3 – $15 ₹274 – ₹1,372
$8 – $50 ₹732 – ₹4,573
10,000
$20 – $200 ₹1,829 – ₹18,292
$30 – $150 ₹2,744 – ₹13,719
$50 – $500 ₹4,573 – ₹45,730
100,000
$200 – $2,000 ₹18,292 – ₹183,200
$300 – $1,500 ₹27,438 – ₹137,190
$500 – $5,000 ₹45,730 – ₹457,300
Pricing References:
Scenario 3: Google Gemini OCR + Google Gemini Mapping
Recommended for: Hindi and mixed-language documents (unified provider)
Documents per month
Google Gemini OCR USD / INR
Google Gemini Mapping USD / INR
Total Monthly Cost USD / INR
1,000
$2 – $20 ₹183 – ₹1,829
$3 – $15 ₹274 – ₹1,372
$8 – $50 ₹732 – ₹4,573
10,000
$20 – $200 ₹1,829 – ₹18,292
$30 – $150 ₹2,744 – ₹13,719
$50 – $500 ₹4,573 – ₹45,730
100,000
$200 – $2,000 ₹18,292 – ₹183,200
$300 – $1,500 ₹27,438 – ₹137,190
$500 – $5,000 ₹45,730 – ₹457,300
Pricing References:
Scenario Comparison Summary
Scenario
OCR Provider
Mapping Provider
Use Case
Estimated Monthly Cost (1,000 documents)
Scenario 1
AWS Textract
AWS Bedrock
English language documents
$8 – $30 ₹732 – ₹2,744
Scenario 2
Google Gemini
AWS Bedrock
Hindi and mixed-language documents
$8 – $50 ₹732 – ₹4,573
Scenario 3
Google Gemini
Google Gemini
Hindi and mixed-language documents (unified provider)
$8 – $50 ₹732 – ₹4,573
Cost Factors
The following factors influence total processing costs:
Document Characteristics
Page count: Multi-page PDFs cost more than single-page images
Document quality: Low-quality scans and complex layouts may require additional processing
Text density: Documents with more text increase token usage for mapping services
Image resolution: Higher resolution images increase token costs for Gemini OCR
Processing Patterns
Language distribution: Higher proportion of Hindi/mixed-language documents increases Gemini usage
Retry scenarios: Network failures and API retries increase total API calls
Mapping complexity: Longer extracted text and larger JSON outputs increase mapping costs
Concurrent processing: High concurrency may affect pricing tiers or require provisioned throughput
Service-Specific Factors
AWS Textract:
Different APIs have different per-page rates
Regional pricing variations
Volume discounts may apply at scale
Google Gemini:
Model version affects pricing (Flash vs Pro)
Token counting varies by content type (text vs images)
Free tier limitations and rate limits
Amazon Bedrock:
Model provider significantly affects pricing
Regional pricing variations
Provisioned throughput options for high-volume use cases
Additional Resources
Cost Estimation Tools
Monitoring and Optimization
AWS Cost Management: Monitor Textract and Bedrock usage through AWS Cost Explorer
Google Cloud Billing: Track Gemini API usage through Google Cloud Console
Best Practices: Implement retry logic, caching, and error handling to minimize unnecessary API calls
Support and Documentation
Important Notes
Pricing Changes: Cloud service pricing is subject to change. Always verify current rates using the official pricing links provided above.
Regional Variations: Pricing may vary significantly by region. Ensure you check pricing for your specific deployment region (
ap-south-1for AWS services).Account Plans: Enterprise agreements, reserved capacity, and volume discounts may affect actual pricing. Contact AWS or Google Cloud sales for custom pricing.
Exchange Rates: INR conversions are approximate and based on current exchange rates. Actual billing will be in USD for AWS services.
Free Tiers: Both AWS and Google Cloud offer free tiers with limitations. Review free tier eligibility and limits before production deployment.
Model Selection: Different models within the same service may have different pricing. Verify pricing for your specific model configuration.
For the most current pricing information, always refer to the official pricing pages linked above.
Last updated