How the check works
A job description is split into sections, then into atomic requirements. Each requirement is classified as a hard gate, core skill, nice-to-have, soft signal, or logistics. The prep score is:
PrepScore = W × Q × Gap
W is the requirement weight (hard gates start at 100; core 60–80; nice-to-have 20–35). Q is the prior probability it gets asked. Gap blends “I have evidence” and “I can explain it.” Tiers are rule-based: a missed hard gate is P0; nice-to-have and soft items cap at P2.
The company check is six heuristic dimensions: stability, growth, management maturity, workload, pay credibility, and hiring sincerity. It is a structured read of the JD, not a due-diligence report. Industry overlays are heuristics too: they add sector signals and a light scoring adjustment, but no facts layer.
Browse the static example reports, the glossary, the FAQ, the comparisons, or the public data. The editorial policy explains authorship and AI disclosure.
Facts are not jargon
Market facts are jurisdiction-level defaults with a source link and a “not legal advice” disclaimer. They are deliberately separate from jargon decoding, which is probabilistic. Country packs currently cover the US, UK, Germany, France, and the Netherlands; the EU pack is a fallback for unclassified euro roles and does not claim country rules.
Fact sources
- GOV.UK · Holiday entitlement, workplace pensions, right to work, employment contracts, IR35
- IRS · 401(k), IRS · independent contractor, USCIS · H-1B
- BUrlG §3, BGB §622, Make it in Germany · Blue Card
- Service-Public · congés payés, période d’essai, mutuelle
- Rijksoverheid · vakantiedagen, proeftijd, business.gov.nl · 30% ruling
Source links were checked on 2026-11-08. Rules change; verify at the source before relying on a fact.
Privacy
No accounts. No cookies. No analytics. No uploads. The JD and your five fields stay in browser memory and disappear on refresh. Opening a fact source is a normal outbound link; nothing is sent by this page.
Limits
- This is a rule-based reading of a JD, not a recruiter’s decision or a legal opinion.
- No auto-apply, no messaging, no scraping, no recruiter personal data.
- Scores are heuristics. Treat them as prompts for questions, not facts.
- Sample companies and roles are fictional.
方法与边界
JD 先分节，再拆成原子要求，分为硬门槛、核心、加分、软性与流程。准备分：PrepScore = W × Q × Gap；W 是要求权重，Q 是被问概率先验，Gap 同时考虑“有没有证据”和“讲不讲得清”。档位由规则决定：硬门槛未满足直接 P0，加分/软性最高只到 P2。
企业体检是六个启发式维度（稳定性、成长性、管理成熟度、工作强度、薪资可信度、招聘诚意），是对 JD 的结构化阅读，不是尽职调查。行业包同样是启发式，只加行业信号和轻量评分，不带事实层。
可继续查看静态示例报告、术语表、常见问题、对比页或公开数据；编辑政策说明内容来源与 AI 使用方式。
市场事实（法定假期、养老金、notice、IR35 等）与“黑话解码”分层：事实带来源链接与“不构成法律建议”提示；黑话是概率判断。当前覆盖 US/UK/DE/FR/NL；EU 仅作为未分类欧元岗位的兜底，不声称国家规则。
隐私：无账号、无 Cookie、无统计、不上传；JD 与五个字段只存在浏览器内存，刷新即失。点击事实来源才会离开本页。
限制：规则版，不是招聘决定或法律意见；不自动投递、不抓取、不采集招聘者信息；分数是提问的起点，不是结论。示例公司与岗位均为虚构。