What a template contains

Each entry is a full prompt body with placeholders such as the language, the framework, or the snippet you want reviewed. You substitute values in the form, preview the assembled text, then copy it into ChatGPT, Claude, Gemini, Copilot Chat or any local model. Nothing is submitted to a model from this page, so the vault costs no tokens to browse.

Why templates beat ad-hoc prompting

A prompt that specifies the role, the constraints, the input format and the expected output shape produces far more stable results than a one-line question, but it is tedious to retype. Keeping the structure fixed and varying only the inputs also makes results comparable when you change models, which is the practical way to tell whether a change in output came from your prompt or from the model.

Related utilities

Pair the vault with the AI and LLM tools to estimate token counts and request costs before you send a long prompt, or browse the full tool catalog.