A Model Context Protocol (MCP) server that enhances Claude AI's capabilities by providing persistent memory storage for Infrastructure-as-Code (IaC) components, with a focus on version tracking and relationship mapping for Terraform and Ansible resources.
Note
This was a personal project to determine the state of AI's ability if the person using it (me) doesn't have subject matter expertise (lack of Python knowledge). Since it has become rather cost prohibitive, I do not intend to develop or maintain this project further.
The IaC Memory MCP Server addresses the challenge of maintaining accurate, version-aware context for IaC components by providing:
- Persistent storage and version tracking for IaC components
- Hierarchical resource organization with URI-based access
- Comprehensive relationship mapping between components
- Version-specific documentation management
- Schema validation and temporal metadata tracking
- Automated relationship analysis and insights
The server implements a sophisticated resource management system with hierarchical URIs:
resources://<platform>/<category>/<name>
Supported platforms:
- terraform
- ansible
- iac (for general infrastructure entities)
Example URIs:
resources://terraform/providers/aws
resources://terraform/resources/aws/s3_bucket
resources://ansible/collections/community.aws
resources://ansible/modules/community.aws/s3_bucket
The server provides dynamic resource templates for standardized access patterns:
- Terraform provider information:
resources://terraform/providers/{provider_name}
- Resource type details:
resources://terraform/resources/{provider_name}/{resource_type}
- Ansible collection data:
resources://ansible/collections/{collection_name}
- Module information:
resources://ansible/modules/{collection_name}/{module_name}
The server implements four specialized prompts for IaC component discovery and analysis:
- Purpose: Search for IaC resources
- Arguments:
provider
: Provider nameresource_type
: Resource type
- Returns: Information about specific resources for the given provider
- Purpose: Analyze an entity and its relationships
- Arguments:
entity_id
: Entity IDinclude_relationships
: Include relationships
- Returns: Detailed entity analysis including name, type, and observations
- Purpose: Get information about a Terraform provider
- Arguments:
provider_name
: Name of the Terraform provider (required)version
: Specific version to query (optional)
- Returns: Detailed provider information for the specified version
- Purpose: Get information about an Ansible module
- Arguments:
collection_name
: Name of the Ansible collection (required)module_name
: Name of the module (required)version
: Specific version to query (optional)
- Returns: Detailed module information for the specified version
The server implements comprehensive tooling for IaC component management:
get_terraform_provider_info
: Retrieve detailed provider information including version and resourceslist_provider_resources
: List all resources available for a specific providerget_terraform_resource_info
: Get detailed information about a specific resource typeadd_terraform_provider
: Register new providers with versioningadd_terraform_resource
: Add resource definitions with schemasupdate_provider_version
: Update provider versions with new documentation
get_ansible_collection_info
: Get detailed information about an Ansible collectionlist_ansible_collections
: List all available Ansible collectionsget_collection_version_history
: View version history of a collectionget_ansible_module_info
: Get detailed information about a specific modulelist_collection_modules
: List all modules in a collectionget_module_version_compatibility
: Check version compatibility of modulesadd_ansible_collection
: Register new Ansible collectionsadd_ansible_module
: Add new modules with validation and documentation
create_entity
: Create new infrastructure entitiesupdate_entity
: Modify existing entity configurationsdelete_entity
: Remove entities with relationship cleanupview_relationships
: Analyze entity dependencies and relationships
The server supports configuration through environment variables:
DATABASE_URL
: SQLite database locationMCP_DEBUG
: Enable debug logging when setMCP_TEST_MODE
: Enable test mode for database resets
For development, create a .env
file:
DATABASE_URL=sqlite:////path/to/db.sqlite
MCP_DEBUG=1
MCP_TEST_MODE=1
"mcpServers": {
"iac-memory": {
"command": "uv",
"args": [
"--directory",
"/path/to/iac-memory-mcp-server",
"run",
"iac-memory-mcp-server"
]
"env": {
"DATABASE_URL": "sqlite:////home/herman/iac.db"
}
}
}
"mcpServers": {
"iac-memory": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/AgentWong/iac-memory-mcp-server.git",
"python",
"-m",
"iac_memory_mcp_server"
],
"env": {
"DATABASE_URL": "sqlite:////home/herman/iac.db"
}
}
}
# Install dependencies
uv sync
# Run tests
uv run pytest
# Development server with MCP Inspector
npx @modelcontextprotocol/inspector uv run iac-memory-mcp-server
This project is licensed under the MIT License - see the LICENSE file for details.