RAG-powered platform that indexes Azure Verified Modules documentation, parses HCL structures, generates embeddings for semantic search, and produces compliant Terraform code with multi-turn dialog support.
Continuously syncs with Microsoft's AVM GitHub repositories to maintain an up-to-date knowledge base.
Fetches module index, READMEs, variables.tf, outputs.tf, and examples from GitHub
Parses Terraform syntax to extract resource definitions, module structures, and dependencies
Splits documentation into context-preserving chunks optimized for retrieval
Generates vector embeddings using text-embedding-004 model for semantic search
Stores embeddings with metadata for fast similarity search and retrieval
Natural language processing pipeline that understands intent, clarifies ambiguity, and generates code.
Extracts resource types, configurations, and dependencies with confidence scoring
Multi-turn dialog to resolve ambiguous requirements before code generation
Semantic search to find relevant AVM documentation and examples
Generates main.tf, variables.tf, outputs.tf with AVM module references
Intelligently merges generated code into existing projects preserving structure
App & PAT authentication
Workspace & run management
Multi-turn dialog to generate AVM-compliant Terraform
Parse HCL, detect deprecated modules, build dependency graph
Requires repository
Inspect generated code with terraform validate/fmt
Push to GitHub and trigger TFC runs
Requires repository
Search the indexed Azure Verified Modules using semantic search
Standalone Mode
Connect a repo for full integration features
"Create an AVM storage account with private endpoint"