Paste any prompt and get a structure score, filler and vague-term counts, a trimmed version with a word-level diff, and the monthly savings.
100% free · no signup · browser-local
Share on XAny prompt: system, user, or a whole template. The linter scores its structure and flags waste.
Eight parts of a brief. Heuristic, not a grade.
Removes filler and collapses whitespace. Review the diff before you ship it.
The linter checks whether the prompt contains the eight parts of a brief, then subtracts for filler phrases and vague terms. Use it as a pre-flight check, not a verdict: a missing part is cheap to fix before the call, expensive after.
Whether the prompt contains the eight parts of a brief — role, task, context, audience, constraints, output format, length cap, and example — plus filler phrases, vague terms, and unresolved placeholders.
No. It removes common filler phrases and collapses whitespace, then shows a word-level diff so you can see exactly what changed. The meaning stays yours; you approve every edit.
No. It is a heuristic that catches a missing brief before you pay for a retry. A prompt can score 100 and still be wrong for your task, and a short prompt can score low but be perfect for a simple job.
Only if the audit still passes. Trimming that drops the score removed something the model needed. Build for correctness first, then cut what you can prove is waste.
The structure score is a heuristic, not a guarantee of output quality. Token counts use the o200k_base BPE tokenizer. Prices are list input rates from the TokenCalc AI catalog.
See how OpenAI, Claude, Llama, and DeepSeek tokenizers split the same prompt.
Strip filler, boilerplate, and duplicate context to cut prompt tokens by up to 35%.
Assemble a structured prompt from 20 original templates, audit the brief, and price it at your volume.
We'll email you when model prices change or new tools launch. Unsubscribe anytime.