A tiny window that rides along in your browser and quietly makes every prompt sharper, using data and rules, not another AI call. Better answers, zero extra tokens, zero extra water.
Most prompt helpers send your text to another model to rewrite it, burning tokens and water to save tokens and water. That is backwards. Ripple uses a deterministic rulebook on your device, so the fix itself costs nothing.
A built-in checklist of what a strong prompt needs. It runs locally, instantly, offline.
No model call means no inference. The improvement step uses 0 tokens and 0 ml of water.
Ripple asks the basics you always forget: role, goal, context, format, length, sources.
Type a lazy prompt. Answer the basics. Watch it become a strong one, with no AI involved.
of water used to improve a prompt. Because the smartest fix is the one that does not need a server at all.
Ripple is an early prototype from Promptd Labs.