Technology Stack

Test Data Workbench is built on a deliberately small, mature set of libraries. Each chapter here explains what a given technology does in the system and why it was chosen over the alternatives, grounded in how the code actually uses it, not a general tutorial.

The stack divides cleanly by responsibility:

Technology	Role in Test Data Workbench
Python	The implementation language; type-hinted throughout.
SQLAlchemy	Schema reflection. Reads the structure the whole system adapts to.
Faker	The value engine behind generated generators, seeded for determinism.
Jinja2	Renders generator source code from templates.
Pydantic	Request and response models for the API layer.
FastAPI	The optional REST interface, with automatic OpenAPI docs.
Rich	The CLI’s tables, panels, and progress output.

• Python
• SQLAlchemy
• Faker
• Jinja2
• Pydantic
• FastAPI
• Rich

A note on what is not here

The predecessor generations of this tool listed Redis and a heavier runtime stack. This generation dropped them. A schema-adaptive generator that emits code does not need a caching layer or a message broker to do its job, and every dependency that does not earn its place is one more thing a user has to install and trust. The short stack is a design choice, not an omission.