Prompt Compressor

Strip redundant filler and whitespace from a prompt to cut token count — see exactly what changed and how much it saves at scale.

Compression Rules

Collapse extra whitespace & blank linesRemove filler phrases ("please note that", "in order to", ...)Remove duplicate consecutive wordsStrip markdown emphasis (**bold**, *italic*) — may reduce clarity for the model

Original Prompt~111 tokens

Compressed Prompt

to complete this task successfully, you should review the attached document very carefully.

 we need this done promptly, and now we don't have a backup plan because the original author is unavailable., summarize the document in three bullet points, and, flag anything that is unique or unusual.

~67 tokens (-39.6%)

Tokens Saved

Per request

Cost Savings at Scale

Requests / day

Estimated monthly input-cost savings

$3.30

Compression is rule-based text editing (whitespace, known filler phrases, duplicate words) — it does not use an LLM and won't rewrite meaning. Always re-read the compressed prompt before shipping it; aggressive phrase removal can occasionally change intent.

About Prompt Compressor

System prompts and templates accumulate filler over time — "please note that," "in order to," redundant whitespace — that costs tokens on every single request without adding meaning. This tool applies a set of rule-based cleanups and shows the token and dollar impact at your actual request volume, so you can decide if the tradeoff is worth it.

Key Features

Toggle individual compression rules (filler phrases, whitespace, duplicate words, markdown stripping).
Live token count and monthly cost savings across GPT, Claude, Gemini, and Llama pricing.
Pairs with the Token Counter and Prompt Differ.