Paste an interview guide, a survey, or the task wording for a usability test. Each question is checked against nine ways phrasing distorts answers — leading, loaded, double-barrelled and the rest — and anything flagged comes back with a rewrite.
Example · one flagged question
“How often do you skip the tutorial?”
LoadedVague quantifier
Assumes the respondent skips the tutorial, which may not be true for them.
“Often” means something different to every respondent, so answers cannot be compared.
Rewrite
“When did you last encounter the tutorial, and what did you do?”
Your questions
One question per line. Numbering is fine, it gets stripped.
Nothing is stored unless you save it afterwards.
The wording is the easy half
Deciding which questions to ask at all — and what you will do with each answer — is where most research goes wrong. That is the part I help teams with.