Kendra

🔍 What is Amazon Kendra?

Amazon Kendra is a highly accurate, intelligent search service powered by machine learning. It enables organizations to search unstructured and structured data across documents, websites, knowledge bases, and more.

✅ Kendra understands natural language questions like “How do I reset my password?” and returns the most relevant documents and direct answers.

---

📚 Use Cases

Industry	Use Case Example
IT Support	Internal documentation search, FAQs
Healthcare	Search medical manuals, policies, compliance
Legal	Discover case laws, compliance docs
Education	Course materials, knowledge base search
Enterprise	Unified search across SharePoint, Confluence, etc

---

🧠 Key Features

Feature	Description
Natural language queries	Understands context and intent
Direct answers	Extracts specific snippets from documents
FAQs support	You can upload CSV/JSON question-answer pairs
Document ranking	Uses ML to rank and score documents by relevance
Synonyms	You can define synonyms (e.g., "laptop" = "notebook")
Access control	Secure document-level filtering with ACLs
Multi-language support	English, French, Spanish, German, Japanese, etc.
Real-time & batch indexing	Automatically sync documents from data sources
Prebuilt connectors	S3, SharePoint, Salesforce, ServiceNow, RDS, Box, etc.

---

🗂️ Supported Data Sources

Source	Notes
Amazon S3	PDFs, DOCX, HTML, TXT, CSV
SharePoint	Online and on-prem
Salesforce	Knowledge base, CRM
ServiceNow	Incident, change, problem KBs
Confluence, Box	Enterprise documentation
RDS / Databases	SQL-based connectors

✅ You can also create custom connectors using Lambda.

---

🧪 Example: Kendra Query API with Python

import boto3

kendra = boto3.client('kendra')

response = kendra.query(
    IndexId='your-index-id',
    QueryText="How do I reset my password?"
)

for result in response['ResultItems']:
    print(result['Type'], "→", result['DocumentTitle']['Text'])
    print("Answer:", result['DocumentExcerpt']['Text'])

---

⚙️ Kendra Architecture Overview

User Query → Kendra API
         ↓
Index (ML-powered search)
         ↓
Connected data sources (S3, SharePoint, Salesforce, etc.)

✅ Kendra supports ACLs, so you can restrict results based on user identity.

---

💰 Pricing Overview (2024)

Kendra Edition	Developer Edition	Enterprise Edition
Ideal for	Testing, small workloads	Production, large-scale
Documents/month	~100K	Up to millions
Queries/month	750/month included	Pay-per-query
Cost	~$810/month	Starts at ~$1,620/month

🧠 Pricing also includes:

• Connector fees (per source)

• API queries beyond free tier

---

🔐 Security & Access

Feature	Support
IAM Policies	✅ Yes
KMS Encryption	✅ For data at rest
Access Control List (ACL)	✅ For user-specific results
VPC Endpoints	✅ Yes (via PrivateLink)

---

🧱 Terraform Support

Terraform support for Amazon Kendra is partial via the aws_kendra_index and aws_kendra_data_source resources.

Create a Kendra Index

resource "aws_kendra_index" "example" {
  name                   = "example-kendra-index"
  role_arn               = aws_iam_role.kendra_role.arn
  edition                = "DEVELOPER_EDITION"

  server_side_encryption_configuration {
    kms_key_id = aws_kms_key.example.arn
  }
}

---

Add a Data Source (e.g., S3)

resource "aws_kendra_data_source" "s3_source" {
  name        = "my-s3-source"
  index_id    = aws_kendra_index.example.id
  type        = "S3"
  role_arn    = aws_iam_role.kendra_role.arn

  configuration {
    s3_configuration {
      bucket_name = "my-kendra-documents"
    }
  }
}

---

✅ TL;DR Summary

Feature	Amazon Kendra
Purpose	Enterprise search across unstructured data
NLP-powered search	✅ Yes
Prebuilt connectors	✅ 10+ supported
Fine-grained access control	✅ Document-level filtering
Supports FAQ	✅ Yes
Pricing	Developer & Enterprise Edition
Languages supported	10+ (incl. EN, FR, DE, ES, JP)
Terraform support	✅ Partial (index, data source)
Free trial	❌ No Free Tier

---