Twelve requirements on the page, only four are real must-haves.

In plain code, not by a model. It reads section headings like 'Requirements' and 'Nice to have', plus per-line phrasing cues such as 'must', 'required', 'preferred', 'bonus', and 'a plus'. The same posting always produces the same split, so you can see how many of the listed requirements are actually load-bearing.

We keep a versioned lexicon of over-broad job-posting phrases: 'wear many hats', 'fast-paced', 'rockstar', 'competitive salary', 'work hard play hard', and more. Each flag explains in plain English what the phrase usually signals and the concrete question to ask. It is a translation, not an accusation.

No, and it never claims to. There are no scam, ghost-job, fake-company, or fraud signals anywhere in this tool. It only decodes the language and the requirements so you can decide for yourself. Treating a vague posting as a fraud would be dishonest, so we do not do it.

When a posting states a salary, we compare its midpoint to the live ScoutJobs band for that role and seniority, built only from employer-stated pay (never predicted). Every comparison carries its sample size and date, and below a data floor we say the band is unknown rather than print a fake-precise number.

From two honest inputs only: how many of the real must-haves your pasted skills cover, and whether the stated pay sits below, in, or above the live band. High fit is apply, partial fit is stretch, low fit is skip. There is no hidden scoring and no pressure, just requirement fit and pay evidence.

Two things. The split, the flags, the pay check, and the verdict all run in plain code, so they are reproducible and never invent a number. And every decode ends with real, currently-open roles from the ScoutJobs index that fit you better than the one you pasted, so you leave with jobs to apply to, not just an opinion.