一天投十个岗位全部石沉大海，多数时候原因不是提案写得差，是投错了单。
而「这单值不值得投」有一半是能机械判定的。把岗位描述整段贴进来（原样贴，套话本身就是信号），代码查六项：对方写了预算吗、交付物吗、技术栈吗、时间吗、业务背景吗、验收标准吗。写得越具体的岗位，发布方越可能是真的准备开工——他已经想过这件事怎么做，也更可能读完你的提案，而且竞争比的是匹配度不是报价。
同时扫七类典型措辞：「long-term partnership」（用未来替代当下的价格，可以相信但不要为它定价）、「simple task」、无偿试做、「unlimited revisions」、极度紧急而没有预算、明说要 best price（在按价格排序的岗位上，写得再好的提案也赢不过报价更低的人）、以及 rockstar／ninja。
还会把你的技能词与岗位描述取交集，告诉你哪几个词必须原样出现在提案前三行——对方写 Next.js 你写「React 全家谱」，在扫读时是不匹配的。
然后是提案结构：第一行证明你读过这个岗位（引用一个具体细节），第二行给一个最相关的证据，第三行提一个显示你想过的问题。提案的目标从来不是成交，是得到一次回复。 而「我是一名有 8 年经验的开发者」这种开头，消耗的是最贵的那一行。
$3/月不限次——一天判十个岗位，值得的可能只有两个。
Ten bids a day into silence usually isn't a writing problem — it's bidding on the wrong jobs.
And half of "is this worth bidding on" is mechanically checkable. Paste the post whole (unedited — the boilerplate is itself a signal) and the code checks six things: did they state a budget, deliverables, a stack, a timeline, business context, acceptance criteria. The more specific the post, the more likely the client is actually ready to start — they've thought about the work, they're more likely to read your proposal, and the competition turns on fit rather than price.
It also scans for seven telltale phrasings: "long-term partnership" (future possibility substituting for present price — believable, but don't price on it), "simple task", unpaid test work, "unlimited revisions", extreme urgency with no budget, explicit requests for the best price (where posts sort by price, a better proposal loses to a cheaper one), and rockstar/ninja.
Then it intersects your skills with the post and tells you which words must appear verbatim in your first three lines — write "React ecosystem" when they wrote Next.js and you don't match on a skim.
Then the structure: line one proves you read this specific post, line two gives your single most relevant proof, line three asks a question that shows you've thought about it. A proposal's goal is never to close — it's to get a reply. And "I am a developer with 8 years of experience" spends the most expensive line in the message.
$3/month, unlimited — you'll screen ten posts a day and maybe two are worth it.
订阅后不限次：完整提案草稿（前三行钩子 + 匹配证据 + 报价锚定 + 结尾提问）、发布方真实痛点与隐藏筛选条件的提取、报价结构建议（区间与条件怎么写），以及回复之后的第二封该说什么。Unlimited runs: a full proposal draft (hook, matched evidence, price anchor, closing question), extraction of the client's real pain and hidden filters, how to structure a range-plus-conditions price, and what the second message should say once they reply.
完整度分数与「该不该投」是本工具的粗口径参考（每写清一项维度记 14 分、每处需留意的措辞扣 12 分），不预测你能不能拿到这单，也不代表发布方的真实意图——写得含糊的好客户和写得完整的坏客户都存在。措辞识别按关键词匹配，会有漏判与误判。本工具不接入任何平台、不代投、不读取你的账号。请勿粘贴发布方的联系方式或平台外交易信息——多数平台禁止绕过平台交易，违反规则可能导致封号。你填写的内容不入库。
The completeness score and the bid/skip call are this tool's rough conventions (14 points per stated dimension, minus 12 per flagged phrase). They do not predict whether you'll win the job and don't reveal the client's intent — vague posts from good clients and polished posts from bad ones both exist. Phrase detection is keyword-based and both misses and over-flags. This tool connects to no platform, submits nothing, and reads no account of yours. Don't paste client contact details or off-platform arrangements — most platforms prohibit circumvention and it can cost you the account. Nothing you enter is stored.