About PromptKit

Most prompt collections are a pile of one-liners with no way to know if any of them actually work. PromptKit is built around evidence instead: every prompt and workflow carries a status, draft, tested, or battle-tested, and battle-tested is a measurement, not a label anyone gets to apply by hand. It has to earn 25+ recorded runs first.

A prompt is useful on its own. copy it, use it, done. A workflow is an ordered set of prompts for an outcome that genuinely needs more than one step, with a guided runner that carries your inputs between steps and tracks which checkpoints passed.

Content is either curated, written and reviewed like code, synced from version-controlled source, or posted by the community and clearly marked as such. Neither is passed off as the other.

PromptKit doesn’t call any LLM on your behalf. The runner is clipboard-driven: it hands you the exact text to paste into whichever model or tool you already use, and tracks your progress locally. No API keys, no usage costs, no vendor lock-in.