Prompt Compressor
A rule-based prompt shortener. It finds verbose patterns ("in order to", "please kindly", "it is important that you"), redundant politeness, doubled spaces, and repeated instructions, then shows the compressed prompt with estimated tokens saved.
Original prompt
Compressed
0
Original tokens (est.)
0
Compressed tokens (est.)
0
Tokens saved (est.)
0%
Reduction
Rules applied
Be careful: meaning must be preserved. These are mechanical rewrites of known filler phrases, plus removal of sentences that appear twice word-for-word. They are safe for most instructions, but always read the compressed version before using it: a phrase like "it is important that you" sometimes carries emphasis you want to keep, and models do respond to emphasis. Token counts here are estimates (characters / 3.6). Shorter prompts are cheaper and often clearer, but never trade away a constraint the model actually needs.