企评社

Interface labels are shown in English. Company reviews, forum posts and tool descriptions are currently available in Chinese only.

Résumé check (score on seven dimensions + in-depth AI review)

Paste text for instant local scoring, or upload a PDF / Word / TXT for server-side text extraction plus an in-depth AI review, commented dimension by dimension with suggestions.

本地诊断:只在你浏览器里分析,不上传、不留存 AI 深度诊断:PDF 浏览器内抽文 / Word·TXT 转换即焚 + 内置大模型,不留存,限 10 次/小时
先说清楚:这是启发式建议,不是专业咨询。 本地引擎只识别「常见模式」(有没有联系方式、有没有写数字、动词强不强),AI 点评也是模型理解下的参考意见; 两者都识别不到行业术语的专业度等隐性质量。把它当检查清单用,别当录用结论。

粘贴简历全文(本地即时诊断)

0 / 20000 字符(不含空白)

输入实时诊断(约 0.15 秒防抖);文本上限 20000 字符(不含空白),超长请先删减旧经历。文本只留在你的浏览器里,本地诊断不上传、不发起任何请求。

AI 深度诊断(服务端 · 限 10 次/小时)

模式:用上面粘贴的文本
本工具需上传文件:仅用于这一次诊断(Word/TXT 文件经服务器转换抽文、转换即焚;PDF 在你浏览器内直接提取文字、不上传文件本身;文字只送内置 AI 分析,不落库、不训练)。 选「用上面粘贴的文本」时,文本同样只送 AI 做这一次分析。AI 临时不可用时,上方的本地诊断结果不受影响。

PDF 在你浏览器内提取文字(不上传文件本身);Word/TXT 上传后转换即焚。AI 诊断与本站其它 AI 工具(如 JD 解码)共享每小时 10 次的频率上限,免费、无需注册;超 20000 字符的文本会截断后送模型(页面会提示)。

一份 HR 眼里「合格」的简历长什么样:七个可检查的维度

联系方式与结构是硬伤,占 45 分。HR 筛简历平均只有几秒:名字下面没有手机号和邮箱,直接翻到下一份;找不到「工作/实习经历」与「教育背景」两个板块,也很少会去猜。页面把这两项拆成 20 分联系方式 + 25 分结构,应届生没有正式工作经历时,项目经历可以部分顶替工作经历的分数,但教育背景不可缺。

篇幅是经验值,不是定论。600 字符以下信息量通常不足,2500 字符以上 HR 平均浏览时间不够、重点被稀释,建议主力经历 600~1500 字符(约一页 A4)。这一维度只占 15 分,页面不会因为你的简历是 1800 字符就把你判死——应届生偏短、资深候选人偏长都正常。

「量化成果」是被低估的一项。「参与系统开发」和「设计并实现支付对账系统,日均处理 50 万笔,差错率 0.01%」的说服力差距,就是量化与没量化的差距。本地引擎统计你简历里条目行中真正出现数字(剔除日期后)的比例:一半以上条目有数字,这一维才拿满分;AI 深度诊断还会逐条点评哪些成果缺数字、怎么补。金额、人数、耗时、百分比、提升幅度,都算。

动词决定简历的「主动感」。「主导 / 设计 / 实现 / 优化 / 落地」开头,读起来是你在做事;「参与 / 协助 / 负责日常」开头,读起来是你在场。本地引擎按词表统计强动词与弱表达的出现次数;AI 深度诊断能识别词表之外的弱表述并给出替换建议——这不是文字洁癖,是筛选效率:HR 靠动词快速判断你的角色深度。

关于本页的数据。本地诊断:简历文本只在你浏览器里被规则引擎读取,页面不发起任何网络请求,不写 localStorage、cookie 或网址参数,刷新即清空。AI 深度诊断:PDF 在你浏览器内直接提取文字(不上传文件本身);Word/TXT 经服务端转换通道抽文(独立容器、转换即焚、不落盘);文字送平台内置大模型分析,文件与文字都不落数据库、不写对象存储、不用于训练,日志只记字符数与耗时,不记内容;AI 临时不可用时保留本地诊断结果。本页不代写、不改写,简历内容始终由你自己掌控。

FAQ

How is the diagnosis done? What is the difference between local and AI?

Two layers. (1) A local rule engine scores instantly on seven dimensions from 0 to 100: contact info (phone/email), structural completeness (work/internship, education), length (600-1500 characters recommended), quantified results (share of bullets containing numbers), wording strength (lead/design/optimize verbs versus participate/assist weak ones), timeline continuity (start-end dates per block), and bullet discipline (one item per line). The text is analyzed only in your browser - no upload, no request at all. (2) The AI deep diagnosis sends the text to the platform's built-in model for per-dimension commentary, an overall assessment and improvement suggestions; PDF text is extracted inside your browser (the file itself is never uploaded), Word/TXT go through the server's text conversion first; scanned or image-only PDFs yield no text - paste the text directly instead. AI results come from model understanding: they are for reference, they are not hiring decisions, and this is not a ghostwriting service.

Which file formats can be uploaded? How is conversion handled?

Supports .pdf / .docx / .doc / .txt up to 20MB per file. PDF text is extracted inside your browser (the file never leaves your device); Word files reuse the site's Document Conversion service to become plain text - done inside an isolated conversion container, the file is deleted the moment conversion finishes (convert-and-incinerate, never stored); TXT is read directly. Scanned or image-only PDFs yield no text - copy and paste the text instead.

Where does my resume data go? Is any of it retained?

Paste mode (local diagnosis): the text is read only by the rule engine in your browser - no upload, no request, refresh clears it. AI deep diagnosis: PDF text is extracted in your browser (the file itself is not uploaded); Word/TXT are used only for this one diagnosis (convert-and-incinerate); the text goes only to the platform's built-in AI model - neither files nor text are written to any database or object storage, and they are not used for training; logs record only character counts and durations, never content. If the AI is temporarily unavailable, the page keeps your local diagnosis results - you are never blocked.

Is there a usage limit? Does it cost anything?

Local diagnosis is unlimited, free, no login. The AI deep diagnosis has rate control: it shares an hourly window with the site's other AI tools (e.g. JD Decoder), capped at 10 per hour - past that you are asked to try later; there is no daily cap, and token usage per call is not capped. Both modes are free and need no registration.

If my score is low, must I rewrite? Can you fix my resume for me?

Not necessarily. The local engine only recognizes common patterns, and the AI commentary is reference opinion under model understanding - neither is a hiring verdict. Fix the hard problems first (contact info, structure, timeline); for the rest, decide against your target role. This page diagnoses and suggests only - it does not ghostwrite or rewrite; every suggestion names what to change and why, and your resume stays yours to control throughout.

Related tools