How techquanta/init-app works
The repository contains two related but distinct products: init-app scaffolds framework-specific projects, while ai-scaffold plans and applies bounded changes to FastAPI projects. The primary flow is CLI input → project configuration and generation → optional health analysis. The AI-native flow validates a plan, previews changes, then applies them with rollback and syntax checks. The graph also preserves documented dbt and local-RAG capabilities. Arrows rely on README statements or SOURCE INDEX references; unsampled implementation details are not inferred.
Main entry points: controller.py Generation controller · ai_scaffold/cli.py AI scaffold CLI · engine/cli.py CLI and UI · planner.py Plan validation
Components 20
Project scaffolding
- CLI and UI
create_app/engine/cli.py - Project specification
create_app/project_spec.py - Generation controller
create_app/initializer/controller.py - Blueprint bundler
create_app/framework/bundler.py - Project generator
create_app/initializer/generator.py - Framework rules
create_app/rules - Path configuration
create_app/path_config.py
Project support
- Project health checks
create_app/doctor.py - dbt adapter support
create_app/dbt_support.py - Local RAG context
create_app/rag_context.py
AI scaffolding
- AI scaffold CLI
ai_scaffold/cli.py - MCP server
ai_scaffold/mcp_server.py
AI planning
- Plan contracts
ai_scaffold/models.py - Plan validation
ai_scaffold/planner.py - Component registry
ai_scaffold/registry.py - Project index
ai_scaffold/context.py - Transactional executor
ai_scaffold/engine.py - Project metadata
ai_scaffold/manifest.py
Other components
- User
- Plan author
Connections 23
- User → CLI and UI: configures project
README.md - CLI and UI → Project specification: loads spec
create_app/engine/cli.py - CLI and UI → Generation controller: starts generation
create_app/engine/cli.py - Generation controller → Path configuration: resolves output
create_app/initializer/controller.py - Generation controller → Blueprint bundler: builds blueprint
create_app/initializer/controller.py - Blueprint bundler → Framework rules: selects rules
create_app/framework/bundler.py - Generation controller → Project generator: renders project
create_app/initializer/controller.py - Generation controller → dbt adapter support: configures dbt
create_app/initializer/controller.py - Generation controller → Local RAG context: writes context bundle
create_app/initializer/controller.py - CLI and UI → Project health checks: runs analysis
create_app/engine/cli.py - User → AI scaffold CLI: submits plan
README.md - Plan author → MCP server: uses MCP tools
ai_scaffold/mcp_server.py - AI scaffold CLI → Plan contracts: parses plans
ai_scaffold/cli.py - AI scaffold CLI → Plan validation: validates plan
ai_scaffold/cli.py - Plan validation → Plan contracts: uses contracts
ai_scaffold/planner.py - Plan validation → Component registry: resolves components
ai_scaffold/planner.py - AI scaffold CLI → Component registry: lists components
ai_scaffold/cli.py - AI scaffold CLI → Transactional executor: previews or applies
README.md - Transactional executor → Project metadata: loads and writes metadata
ai_scaffold/engine.py - AI scaffold CLI → Project index: builds project index
ai_scaffold/cli.py - Project index → Project metadata: loads metadata
ai_scaffold/context.py - MCP server → Plan validation: exposes planning
ai_scaffold/mcp_server.py - MCP server → Transactional executor: exposes application
ai_scaffold/mcp_server.py