Github Awslabs Mcp Official Mcp Servers For Aws Github

Loading... Loading... Menu - .github/CODEOWNERS - docusaurus/docs/servers/sagemaker-unified-studio-spark-troubleshooting-mcp-server.md - docusaurus/docs/servers/sagemaker-unified-studio-spark-upgrade-mcp-server.md - docusaurus/sidebars.ts - docusaurus/static/assets/server-cards.json - src/aws-bedrock-data-automation-mcp-server/README.md - src/aws-diagram-mcp-server/README.md - src/aws-documentation-mcp-server/README.md - src/aws-support-mcp-server/README.md - src/bedrock-kb-retrieval-mcp-server/README.md - src/cdk-mcp-server/README.md - src/cfn-mcp-server/README.md - src/core-mcp-server/README.md - src/frontend-mcp-server/README.md - src/lambda-tool-mcp-server/README.md - src/mysql-mcp-server/README.md - src/nova-canvas-mcp-server/README.md - src/postgres-mcp-server/awslabs/postgres_mcp_server/connection/cp_api_connection.py - src/postgres-mcp-server/awslabs/postgres_mcp_server/connection/db_connection_map.py - src/postgres-mcp-server/kiro_power/POWER.md - src/postgres-mcp-server/kiro_power/steering/aurora-postgres-mcp.md - src/postgres-mcp-server/kiro_power/steering/aurora-postgres.md - src/postgres-mcp-server/kiro_proj_steering/tech.md - src/sagemaker-unified-studio-spark-troubleshooting-mcp-server/README.md - src/sagemaker-unified-studio-spark-upgrade-mcp-server/README.md - src/terraform-mcp-server/README.md This page provides a comprehensive catalog of all MCP servers available in the awslabs/mcp repository.

Each server is a standalone Python package that enables AI assistants to interact with specific AWS services or perform specialized tasks. For information about the Model Context Protocol itself and how clients connect to servers, see What is Model Context Protocol?. For understanding how servers fit into different development workflows, see Server Workflow Classifications. The repository contains 50+ specialized MCP servers organized into 9 functional categories. Each server is independently versioned and published to PyPI, allowing users to install only the servers they need.

Sources: docusaurus/static/assets/server-cards.json1-2048 .github/CODEOWNERS1-95 All MCP servers are located under the src/ directory, with each server in its own subdirectory.

The structure follows a consistent pattern: src/ ├── aws-api-mcp-server/ # Core AWS CLI integration ├── core-mcp-server/ # Dynamic proxy and orchestration ├── bedrock-kb-retrieval-mcp-server/ # Knowledge base retrieval ├── dynamodb-mcp-server/ # NoSQL database operations └── [50+ other servers...] Each server directory contains: pyproject.toml - Package metadata and dependenciesREADME.md - Usage documentation and examplesawslabs/[server-name]/ - Python package codetests/ - Test suite (optional)Dockerfile - Container build instructions (optional) Sources: .github/CODEOWNERS19-88 docusaurus/static/assets/server-cards.json58-2048 Servers are organized into 9 functional categories with associated icons and descriptions: Sources: docusaurus/static/assets/server-cards.json2-56 Location: src/aws-api-mcp-server/ The foundational server providing secure AWS CLI command execution with parsing, validation, and consent mechanisms.

Supports 350+ AWS services through the AWS CLI. Key Features: - Command parsing and validation via `awslabs.aws_api_mcp_server.parser` - Security policy enforcement via `mcp-security-policy.json` - Consent elicitation for mutating operations - Read-only mode support - Custom validator framework Sources: docusaurus/static/assets/server-cards.json101-119 .github/CODEOWNERS30 Location: src/core-mcp-server/ Dynamic proxy server that loads 46+ specialized servers based on role-based environment variables. Implements a unified entry point for AWS development workflows.

Key Features: - Role-based server loading (17 predefined roles) - Planning and orchestration via prompt_understanding tool - Eliminates manual server configuration - Deduplicates server imports across roles Role Examples: aws-foundation →aws-knowledge-server ,aws-api-server solutions-architect →diagram-server ,pricing-server ,cost-explorer-server sql-db-specialist →postgres-server ,mysql-server ,aurora-dsql-server Sources: src/core-mcp-server/README.md1-160 docusaurus/static/assets/server-cards.json102-119 Location: src/aws-documentation-mcp-server/ Provides real-time access to AWS documentation with markdown conversion, search capabilities, and content recommendations.

Available Tools: read_documentation - Fetches and converts AWS docs to markdownsearch_documentation - Official AWS Documentation Search API (global only)recommend - Content recommendations for doc pages (global only)get_available_services - Lists AWS services in China regions (China only) Configuration: AWS_DOCUMENTATION_PARTITION - Set toaws oraws-cn MCP_USER_AGENT - Custom User-Agent for corporate proxies Sources: src/aws-documentation-mcp-server/README.md1-201 docusaurus/static/assets/server-cards.json58-79 Location: src/aws-knowledge-mcp-server/ Retrieves latest AWS documentation, code samples, and official content through a real-time search API with citation support.

Sources: docusaurus/static/assets/server-cards.json80-100 .github/CODEOWNERS40 Sources: docusaurus/static/assets/server-cards.json120-322 docusaurus/sidebars.ts42-58 Location: src/aws-iac-mcp-server/ Comprehensive Infrastructure as Code server with CloudFormation Guard rules, policy validation, and IaC best practices. Replaces the deprecated cdk-mcp-server . Sources: .github/CODEOWNERS38 docusaurus/sidebars.ts45 Location: src/terraform-mcp-server/ Terraform workflows with integrated Checkov security scanning, AWS provider documentation search, and module analysis.

Available Tools: RunTerraformCommand - Execute terraform init/plan/validate/apply/destroyRunTerragruntCommand - Execute terragrunt operationsRunCheckovScan - Security and compliance scanningSearchAWSProviderResource - Find AWS/AWSCC provider resourcesAnalyzeTerraformModule - Extract module variables and outputs Sources: src/terraform-mcp-server/README.md1-151 docusaurus/static/assets/server-cards.json182-201 Location: src/cfn-mcp-server/ Direct CloudFormation resource management via Cloud Control API with support for 1100+ AWS resource types.

Available Tools: create_resource - Declarative resource creationget_resource - Retrieve resource detailsupdate_resource - Declarative updatesdelete_resource - Resource deletionlist_resources - Enumerate resources by typeget_resource_schema_information - CloudFormation schema retrievalcreate_template - Generate IaC templates from resources Read-Only Mode: Add --readonly flag to prevent mutations Sources: src/cfn-mcp-server/README.md1-241 docusaurus/static/assets/server-cards.json143-160 Location: src/ecs-mcp-server/ Container orchestration and ECS application deployment with Express Mode for rapid development and AWS Knowledge Proxy integration. Sources: .github/CODEOWNERS62 docusaurus/static/assets/server-cards.json224-242 Location: src/lambda-tool-mcp-server/ Bridge between MCP clients and AWS Lambda functions, enabling AI models to invoke Lambda functions as tools without code changes.

Supports EventBridge Schema Registry integration for formal JSON Schema validation. Function Selection: FUNCTION_PREFIX - Filter by function name prefixFUNCTION_LIST - Comma-separated list of specific functionsFUNCTION_TAG_KEY +FUNCTION_TAG_VALUE - Filter by tagsFUNCTION_INPUT_SCHEMA_ARN_TAG_KEY - Tag for schema ARN reference Architecture Pattern: Segregation of duties - model invokes functions, functions access AWS services via function role Sources: src/lambda-tool-mcp-server/README.md1-273 docusaurus/static/assets/server-cards.json284-303 Location: src/bedrock-kb-retrieval-mcp-server/ Query Amazon Bedrock Knowledge Bases with reranking, filtering, and multi-source query capabilities.

Available Tools: - Discover knowledge bases by name or tag - Query with natural language - Filter results by data source (include/exclude/prioritize) - Rerank results using Bedrock reranking models Configuration: KB_INCLUSION_TAG_KEY - Filter KBs by tagBEDROCK_KB_RERANKING_ENABLED - Global reranking toggle (default:false ) Limitations: IMAGE content types excluded, reranking requires additional permissions Sources: src/bedrock-kb-retrieval-mcp-server/README.md1-177 docusaurus/static/assets/server-cards.json324-343 Location: src/nova-canvas-mcp-server/ AI image generation using Amazon Nova Canvas with text and color-guided generation.

Available Tools: generate_image - Create images from text promptsgenerate_image_with_colors - Color palette-guided generation (up to 10 hex colors) Parameters: - Dimensions: 320-4096px - Multiple images: 1-5 per request - cfg_scale: 1.1-10.0 - Seeded generation support - Negative prompting Sources: src/nova-canvas-mcp-server/README.md1-150 docusaurus/static/assets/server-cards.json430-449 Location: src/aws-bedrock-data-automation-mcp-server/ Analyze documents, images, videos, and audio files using Amazon Bedrock Data Automation projects.

Available Tools: getprojects - List data automation projectsgetprojectdetails - Get project details by ARNanalyzeasset - Extract insights from unstructured content Environment Variables: AWS_BUCKET_NAME - S3 bucket for assets/resultsBASE_DIR - Base directory for file operations Sources: src/aws-bedrock-data-automation-mcp-server/README.md1-202 docusaurus/static/assets/server-cards.json451-470 Location: src/sagemaker-ai-mcp-server/ SageMaker HyperPod cluster management and model development workflows. Sources: .github/CODEOWNERS80 docusaurus/static/assets/server-cards.json494-516 Sources: docusaurus/static/assets/server-cards.json517-809 docusaurus/sidebars.ts75-96 Location: src/dynamodb-mcp-server/ Complete DynamoDB operations with data modeling workflow, validation using DynamoDB Local, and source database analyzer.

Key Features: - Table management and CRUD operations - Data modeling guidance - Local validation before deployment - Source database analysis for migration planning Sources: .github/CODEOWNERS61 docusaurus/static/assets/server-cards.json517-537 Location: src/postgres-mcp-server/ Aurora PostgreSQL operations via RDS Data API with multi-connection method support and Power system guidance.

Connection Methods: rdsapi - RDS Data API (serverless, no network connectivity required)pgwire - Direct PostgreSQL wire protocolpgwire_iam - PostgreSQL with IAM authentication Available Tools: mcp_postgres_create_cluster - Async cluster creation (returns job_id)mcp_postgres_get_job_status - Monitor async operationsmcp_postgres_connect_to_database - Establish connectionmcp_postgres_run_query - Execute SQL queriesmcp_postgres_get_table_schema - Schema explorationmcp_postgres_get_database_connection_info - List active connections Power System: Includes comprehensive guidance via src/postgres-mcp-server/kiro_power/: POWER.md - Overview and MCP integration patternssteering/aurora-postgres-mcp.md - MCP tool usage policiessteering/aurora-postgres.md - Complete development guide (400+ lines) Safety Features: - Read-only by default (requires --allow_write_query flag) - Query policies via steering files - Cluster creation monitoring (poll every 30-60 seconds) Sources: src/postgres-mcp-server/README.md1-201 src/postgres-mcp-server/kiro_power/POWER.md1-308 src/postgres-mcp-server/kiro_power/steering/aurora-postgres-mcp.md1-50 Location: src/mysql-mcp-server/ Aurora MySQL and RDS MySQL operations via RDS Data API with support for direct MySQL connection method.

Connection Methods: - RDS Data API (recommended for Aurora) - Direct MySQL connection via asyncmy (for Aurora, RDS, MariaDB, self-hosted) Configuration: --resource_arn - For RDS Data API connection--hostname - For direct MySQL connection--secret_arn - AWS Secrets Manager secret ARN--readonly - Enable read-only mode (default:True ) Sources: src/mysql-mcp-server/README.md1-172 .github/CODEOWNERS73 Location: src/s3-tables-mcp-server/ Manage, query, and ingest S3-based Iceberg tables with SQL parsing, CSV-to-table conversion, and metadata discovery.

Key Features: - SQL query execution against Iceberg tables - CSV ingestion to table buckets - Metadata discovery - Lazy tag loading for optimized file search Sources: .github/CODEOWNERS79 docusaurus/static/assets/server-cards.json559-581 Location: src/aws-diagram-mcp-server/ Generate architecture diagrams using Python diagrams DSL with integrated security code scanning.

Supported Diagram Types: - AWS architecture diagrams - Sequence diagrams - Flow charts - Class diagrams Features: - GraphViz-based rendering - Security code scanning before execution - Customizable appearance and layout Prerequisites: Requires GraphViz installation Sources: src/aws-diagram-mcp-server/README.md1-153 docusaurus/static/assets/server-cards.json852-871 Location: src/git-repo-research-mcp-server/ Semantic code search and repository analysis for understanding codebases. Sources: .github/CODEOWNERS67 docusaurus/static/assets/server-cards.json812-830 Location: src/code-doc-gen-mcp-server/ Automated documentation generation from code analysis. Sources: .github/CODEOWNERS56 docusaurus/static/assets/server-cards.json832-850 Location: src/frontend-mcp-server/ React and modern web development guidance with AWS Amplify integration.

Available Tool: GetReactDocsByTopic - Comprehensive documentation on React topics Topics: - Essential Knowledge - Fundamental React concepts - Basic UI Setup - Tailwind CSS and shadcn/ui - Authentication - AWS Amplify auth integration - Routing - React Router implementation - Customizing - Theming with Amplify components - Creating Components - React components with AWS - Troubleshooting - Common issues and solutions Sources: src/frontend-mcp-server/README.md1-97 docusaurus/static/assets/server-cards.json872-892 Location: src/amazon-sns-sqs-mcp-server/ Event-driven messaging and queue management with SNS and SQS integration. Sources: .github/CODEOWNERS28 docusaurus/sidebars.ts122-123 Location: src/stepfunctions-tool-mcp-server/ AWS Step Functions workflow orchestration and execution.

Sources: .github/CODEOWNERS83 docusaurus/sidebars.ts54 Location: src/aws-pricing-mcp-server/ Pre-deployment cost estimates and pricing information for AWS services. Sources: .github/CODEOWNERS44 docusaurus/sidebars.ts129 Location: src/cost-explorer-mcp-server/ AWS Cost Explorer integration for cost analysis and reporting. Sources: .github/CODEOWNERS58 docusaurus/sidebars.ts130 Location: src/cloudwatch-mcp-server/ Amazon CloudWatch integration for metrics, logs, and alarms management. Sources: .github/CODEOWNERS55 docusaurus/sidebars.ts131 Location: src/aws-support-mcp-server/ AWS Support case management and operations.

Available Tools: - Create and manage support cases - Retrieve case information and communications - Add communications to cases - Resolve support cases - Determine appropriate Issue Type, Service Code, Category Code - Determine appropriate Severity Level Requirements: - Business, Enterprise On-Ramp, or Enterprise Support plan - IAM permissions for Support API access Sources: src/aws-support-mcp-server/README.md1-142 docusaurus/static/assets/server-cards.json306-322 Location: src/well-architected-security-mcp-server/ AWS Well-Architected Security Pillar compliance checks and recommendations. Sources: .github/CODEOWNERS88 docusaurus/sidebars.ts133 Location: src/aws-healthomics-mcp-server/ Comprehensive AWS HealthOmics integration for genomics workflows.

Key Features: - Workflow management (WDL/CWL) - Run execution and monitoring - Genomics file search engine - Performance analysis Sources: .github/CODEOWNERS37 docusaurus/sidebars.ts142 Two servers are fully managed AWS services accessed via HTTPS with SigV4 authentication through mcp-proxy-for-aws These require the MCP Proxy for AWS to handle secure communication. Sources: src/sagemaker-unified-studio-spark-troubleshooting-mcp-server/README.md1-240 src/sagemaker-unified-studio-spark-upgrade-mcp-server/README.md1-172 Endpoint: https://sagemaker-unified-studio-mcp.{region}.api.aws/spark-troubleshooting/mcp Troubleshoot Apache Spark applications on Amazon EMR, AWS Glue, and SageMaker Notebooks.

Key Features: - Intelligent failure analysis via Spark event logs - Multi-platform support (PySpark and Scala) - Automated feature extraction from Spark History Server - GenAI root cause analysis - Code recommendation engine Supported Platforms: - Amazon EMR on EC2 (via EMR Persistent UI) - Amazon EMR Serverless (via Spark History Server) - AWS Glue (via Glue Studio Spark UI) - Amazon SageMaker Notebooks Companion Server: spark-code-recommendation at /spark-code-recommendation/mcp endpoint Configuration: Sources: src/sagemaker-unified-studio-spark-troubleshooting-mcp-server/README.md1-240 .github/CODEOWNERS82 Endpoint: https://sagemaker-unified-studio-mcp.{region}.api.aws/spark-upgrade/mcp Accelerate Apache Spark version upgrades from 2.4 to 3.5 on Amazon EMR.

Key Features: - Project analysis and upgrade planning - Automated code transformation (PySpark and Scala) - Dependency and build management (Maven/SBT/pip) - Comprehensive testing and validation - Data quality validation - EMR integration and monitoring Upgrade Paths: - EMR-EC2: 5.20.0+ → 7.12.0 - EMR-Serverless: 6.6.0+ → 7.12.0 Workflow: - Planning - Project structure analysis - Compile & Build - Environment updates, iterative fixes - Spark Code Edit - Version incompatibility resolution - Execution & Validation - Remote EMR job submission - Observability - Progress tracking and analysis Sources: src/sagemaker-unified-studio-spark-upgrade-mcp-server/README.md1-172 .github/CODEOWNERS81 All locally-run servers follow this pattern: Sources: src/aws-documentation-mcp-server/README.md27-45 src/bedrock-kb-retrieval-mcp-server/README.md82-102 Windows requires explicit uv tool run syntax: Sources: src/aws-documentation-mcp-server/README.md72-98 src/nova-canvas-mcp-server/README.md64-91 After building with docker build -t awslabs/{server-name} .

Sources: src/bedrock-kb-retrieval-mcp-server/README.md133-169 src/nova-canvas-mcp-server/README.md94-126 Each server maintains independent semantic versioning. Current versions are tracked in server-cards.json: Releases are automated via .github/workflows/release.yml with PyPI and npmjs publishing.

Sources: docusaurus/static/assets/server-cards.json1-2048 High-Level Diagram 3 Sources: High-Level Diagram 4, src/core-mcp-server/README.md1-160 Building Infrastructure: - Start with aws-iac-mcp-server for CloudFormation Guard validation - Use terraform-mcp-server for Terraform workflows with Checkov - Use ccapi-mcp-server for direct Cloud Control API access Working with Databases: - SQL: postgres-mcp-server ,mysql-mcp-server ,aurora-dsql-mcp-server - NoSQL: dynamodb-mcp-server ,documentdb-mcp-server ,neptune-mcp-server - Caching: elasticache-mcp-server ,valkey-mcp-server ,memcached-mcp-server AI/ML Development: - Knowledge retrieval: bedrock-kb-retrieval-mcp-server - Image generation: nova-canvas-mcp-server - ML operations: sagemaker-ai-mcp-server - Document analysis: aws-bedrock-data-automation-mcp-server Cost Management: - Pre-deployment estimates: aws-pricing-mcp-server - Cost analysis: cost-explorer-mcp-server - Monitoring: cloudwatch-mcp-server Development Tools: - Documentation: aws-documentation-mcp-server ,aws-knowledge-mcp-server - Code analysis: git-repo-research-mcp-server ,code-doc-gen-mcp-server - Architecture: aws-diagram-mcp-server Sources: docusaurus/static/assets/server-cards.json1-2048 src/core-mcp-server/README.md20-41 Servers are tagged with workflow patterns in server-cards.json: - vibe-coding - Interactive development with immediate feedback (26 servers) - conversational - Natural language queries and responses (18 servers) - autonomous - Long-running operations with minimal interaction (16 servers) See Server Workflow Classifications for detailed workflow explanations.

Sources: docusaurus/static/assets/server-cards.json74-78 docusaurus/static/assets/server-cards.json116-118 All servers requiring AWS access support: AWS_PROFILE - Named profile from~/.aws/credentials AWS_REGION - Target AWS region- Environment variables: AWS_ACCESS_KEY_ID ,AWS_SECRET_ACCESS_KEY ,AWS_SESSION_TOKEN Sources: src/nova-canvas-mcp-server/README.md138-150 src/postgres-mcp-server/kiro_proj_steering/tech.md24-33 Standard logging environment variable across all servers: FASTMCP_LOG_LEVEL - Values:DEBUG ,INFO ,WARNING ,ERROR ,CRITICAL Sources: src/aws-documentation-mcp-server/README.md132-137 Many servers support read-only mode: --readonly flag orREAD_OPERATIONS_ONLY environment variableREQUIRE_MUTATION_CONSENT - Elicit user consent for mutationsFILE_ACCESS_MODE - Control file system access (NO_ACCESS ,WORKDIR ,UNRESTRICTED ) Sources: src/cfn-mcp-server/README.md50-70 High-Level Diagram 5 Each server's README contains: - Detailed feature descriptions - Installation instructions (uvx, Windows, Docker) - Tool/resource documentation - Configuration options - Usage examples - Security considerations Server READMEs are located at src/{server-name}/README.md For architectural patterns and security models, see: - System Design - Monorepo structure and namespace packages - Security & Permission Model - Multi-layer security architecture - Package Structure & Dependencies - Common dependency patterns Sources: docusaurus/static/assets/server-cards.json58-2048 .github/CODEOWNERS19-88 Refresh this wiki - Available MCP Servers - Purpose and Scope - Server Organization - Directory Structure - Category System - Category Breakdown - Core Servers (⚡) - aws-api-mcp-server - core-mcp-server - Documentation Servers (📚) - aws-documentation-mcp-server - aws-knowledge-mcp-server - Infrastructure & Deployment Servers (🏗️) - aws-iac-mcp-server - terraform-mcp-server - cfn-mcp-server - ecs-mcp-server - lambda-tool-mcp-server - AI & Machine Learning Servers (🤖) - bedrock-kb-retrieval-mcp-server - nova-canvas-mcp-server - aws-bedrock-data-automation-mcp-server - sagemaker-ai-mcp-server - Data & Analytics Servers (📊) - dynamodb-mcp-server - postgres-mcp-server - mysql-mcp-server - s3-tables-mcp-server - Developer Tools & Support Servers (🛠️) - aws-diagram-mcp-server - git-repo-research-mcp-server - code-doc-gen-mcp-server - frontend-mcp-server - Integration & Messaging Servers (📡) - amazon-sns-sqs-mcp-server - stepfunctions-tool-mcp-server - Cost & Operations Servers (💰) - aws-pricing-mcp-server - cost-explorer-mcp-server - cloudwatch-mcp-server - aws-support-mcp-server - well-architected-security-mcp-server - Healthcare & Lifesciences Servers (🧠) - aws-healthomics-mcp-server - Remote MCP Servers (AWS-Managed) - sagemaker-unified-studio-spark-troubleshooting-mcp-server - sagemaker-unified-studio-spark-upgrade-mcp-server - Server Installation Patterns - Standard Installation via uvx - Windows Installation - Docker Installation - Server Versioning and Releases - Server Dependency Graph - Server Discovery and Selection - By Use Case - By Workflow Classification - Common Configuration Patterns - AWS Credentials - Logging Configuration - Security Configuration - Finding More Information

People Also Asked

• GitHub - awslabs/mcp: Official MCP Servers for AWS
• Welcome to Open Source MCP Servers for AWS - awslabs.github.io
• AWS MCP Servers | Awesome MCP Servers
• AWS MCP Server Documentation
• Available MCP Servers | awslabs/mcp | DeepWiki
• Open source MCP servers for AWS | Welcome to Open Source MCP Servers ...
• GitHub - SherineRizk/awslabs-mcp: Official MCP Servers for AWS