Check your job description for unconscious bias.

Paste your draft JD. Get an instant report on gendered language, age-coded phrasing, ableist defaults, "rockstar"-style buzzwords, and vague red-flag phrases. Free. Private (runs in your browser — your JD never leaves your device). No signup. No LLM.

How to read the results: this tool surfaces empirically-studied language patterns. It doesn't judge intent. Each flag explains why the word is in the list — you decide whether to keep it, replace it, or ignore the warning. The job is to inform, not to mandate.

What this tool catches

Six categories of bias signal, drawn from peer-reviewed research and open-source word lists:

  • Gendered language. The Gaucher, Friesen & Kay (2011) research found that masculine-coded and feminine-coded words in job ads predictably affect application rates by gender. Words like "competitive", "dominant", "independent" are masculine-coded; "supportive", "collaborative", "compassionate" are feminine-coded. The tool counts both lists and tells you which way your JD leans.
  • Age-coded phrasing. Phrases like "digital native", "recent graduate", "young", "energetic" can correlate with age discrimination claims. In several jurisdictions (US ADEA, EU directives) age-coded language in job ads can trigger legal exposure.
  • Ableist defaults. Physical-ability requirements without job justification ("must be able to stand for 8 hours" in a desk job) plus ableist metaphors ("see eye to eye", "tone-deaf"). These can exclude qualified disabled candidates.
  • Bro-culture buzzwords. "Rockstar", "ninja", "guru", "wizard", "killer" — HR research shows these words measurably reduce female and older-candidate application rates without adding any real information about the role.
  • Vague red-flag phrases. "Self-starter", "fast-paced", "wear many hats", "like a family", "work hard play hard" — these correlate with poor work-life-balance reviews on Glassdoor/Indeed. They might be honest, or might be sending signals you didn't intend.
  • Educational gatekeeping. "Ivy League", "top-tier university", "prestigious university" exclude qualified candidates from non-elite institutions. Unless the role legally requires specific accreditation, this is signalling not screening.

Rules and citations

Gender-coded word lists — source

The masculine-coded and feminine-coded word stems come from:

Gaucher, D., Friesen, J., & Kay, A. C. (2011). "Evidence that gendered wording in job advertisements exists and sustains gender inequality". Journal of Personality and Social Psychology, 101(1), 109–128.

As implemented in Kat Matfield's open-source Gender Decoder (MIT licensed). We use word stems (e.g. "analy" matches "analyze", "analysis", "analytical") and present the result as a balance, not as individual judgments.

Age-coded phrases

Drawn from subsequent HR-research surveys on age discrimination in job advertisements. Critical-severity terms ("digital native", "recent graduate", "young") have appeared in actual EEOC age-discrimination cases. Warning-severity terms ("energetic", "vibrant", "dynamic") are more contextual but recurringly flagged in DEI training materials.

Ableist defaults and metaphors

Physical-ability requirements that lack job justification create unnecessary barriers. Even ostensibly inclusive phrasing like "must be able to stand" excludes qualified seated workers if the role doesn't actually require standing. Ableist metaphors ("see eye to eye", "tone-deaf") are everyday phrases but their accumulation signals an exclusionary norm.

Bro-culture buzzwords

Words like "rockstar", "ninja", "guru", "wizard" have been shown in multiple HR studies to reduce application rates from women and older candidates without conveying actual role information. They're descriptively meaningless ("a wizard at what, exactly?") and signal a culture rather than describe a job.

Vague red-flag phrases

"Self-starter", "fast-paced", "wear many hats", "work hard play hard", "like a family" — these phrases correlate with negative work-life-balance reviews on Glassdoor and Indeed. They might be accurate descriptions of your culture (in which case flagging them is informational), or they might be unintended signals. Either way, surfacing them helps the author ask whether the phrasing is doing the work intended.

Privacy

Your job description never leaves your browser. The entire analysis runs as JavaScript on your device. You can verify this by opening your browser's developer tools (Network tab) and clicking Analyze — you'll see zero outbound requests. The host (Cloudflare Pages) sees anonymous page views, nothing else.

Why this exists

Textio is the dominant paid SaaS in this space, but enterprise pricing starts around $15k/year, it's AI-based, and it requires you to paste your JD into their service. That's prohibitive for SMB HR, solo recruiters, hiring managers, and anyone whose legal team can't sign off on pasting internal drafts into a third-party AI service.

This tool sits in the gap: free, private, deterministic, no signup. Made by LanceLabs. Feedback welcome: what category did we miss? What flag is wrong?

Helped your last hire feel less awkward to write? Keep this tool ad-free.

No signups, no popups, no "upgrade to export" — this tool is free because some folks support the work. If it helped you, a coffee keeps it going.

Buy me a coffee — ko-fi.com/lancelabs