Gendered wording check
Paste a job ad, a job detail page or a piece of social copy. See whether the wording leans masculine-coded, leans feminine-coded, or reads as neutral, then get plain swaps to even it out.
Swap these for neutral alternatives
Starting points, not rules. Context matters, and you don't need to change every word.
Your copy, with coded words marked
The decoder scans your text for words that research has linked to a masculine or feminine slant. It counts how many of each it finds, tells you which way the copy leans overall, and suggests neutral swaps. The point isn't to police language, it's to show you what signal your wording is sending so you can match it to the people you want to reach.
The tool uses a fixed list of word stems drawn from the study below. It matches the start of each word, so lead catches "lead", "leader" and "leadership", and support catches "support", "supportive" and "supporting".
No. The words aren't describing applicants, and the research makes no claim about what anyone can or can't do. It's about the word as a signal, not the trait.
Decades of cultural association have tied words like "competitive", "dominant" and "independent" to a male stereotype. When a lot of them stack up in one ad, they act as a subtle belonging cue. In the studies, women were more likely to read those roles as a poorer fit, a quiet sense of "people like me don't really work here", and rated them less appealing, regardless of how independent or competitive they actually were. The effect ran through belonging, not perceived ability. And tellingly, not one participant across the five studies consciously noticed the wording.
So a masculine-coded ad isn't insulting anyone. It's narrowing your audience by accident. Swapping a few words doesn't water the role down, it removes a signal that was filtering people out before they'd even reached the responsibilities.
It's a fair challenge. The tool isn't deciding that "caring" is female or "leader" is male. It's reflecting associations that already sit in the language, which the research measured by asking people which traits they link to each gender. Naming the pattern is what lets you decide whether to keep it.
It's also a deliberately blunt model. It reads in two directions, masculine and feminine, because that's how the original research coded the words. That binary doesn't capture how gender actually works and it won't speak to everyone, including non-binary readers. So treat the result as a read on stereotyped wording, not a statement about anyone's identity.
The meter weighs how balanced the two sides are. A roughly even mix reads as neutral, a clear majority reads as a lean, and a heavy majority reads as strongly coded. It's the balance that counts, so a long ad isn't marked down just for using more words.
The evidence isn't symmetrical. In the research, masculine wording lowered a job's appeal for women and their sense that they'd belong, and it made even traditionally female roles more appealing to men. Feminine wording didn't push men away in the same way. So if your aim is the widest pool, neutral or lightly balanced copy tends to serve you best. If you're deliberately speaking to one audience, the meter just tells you whether your wording backs that up.