Four steps, in order, for every request that asks for a language other than your own.
Your domain points at the translation proxy instead of straight at your server. A request for /es/pricing arrives at the proxy, not at your origin.
The proxy strips the language prefix and requests /pricing from your own server, exactly as a browser would. Your site is unchanged and does not know the difference.
可见文本、链接标题、表单标签、Alt text 以及元标题和描述都会从 HTML 中提取,并在完整的页面上下文中进行翻译,从而确保单词的翻译方式符合其在句子中的用法。
代理将翻译内容写回原 HTML 并返回。结果会缓存在边缘节点,因此下一位访客请求该语言的页面时,无需再次进行翻译工作即可直接获取。
Every architecture trades something away. These are the trades a proxy makes, including ours.
无需在您的代码库或内容管理系统中安装任何东西。这正是代理存在的意义:没有工程时间的团队,或是不想改动现有技术栈的团队,依然可以发布多语言网站。
翻译发生在页面到达浏览器之前,因此爬虫接收到的是目标语言的完整渲染页面,而不是一个稍后由脚本重写的英文页面。
Translated pages are served from your own domain under a path such as /es/ or /ja-jp/, so each language inherits the authority the domain has already earned.
A page published this morning is translated the first time somebody requests it. There is no export, no re-import, and no build step to re-run.
全代理模式意味着您的域名解析到我们。这是一种真正的依赖,这也是我们同时提供两种更轻量模式的原因。如果您尚未准备好,混合模式可以在保持 DNS 不变的情况下,为爬虫提供服务器端渲染的翻译。
代理会读取 HTML,而绘制在 JPEG 中的文本并非 HTML。Alt text 会自动翻译。嵌入在图像本身的文本仅针对您选择开启的图像进行翻译,翻译后的图像会针对该语言进行缓存。
Pages a crawler can never reach still translate for the visitor who is signed in, but they cannot be pre-translated by crawling alone. Authenticated pages are discovered separately.
代理进行持续翻译,因此按持续使用量计费。我们不会对您已付费的句子重复计费:翻译后的单词会永久存储,而不会每月再次计费。
Most translation services give you one install mode and expect you to live with its trade-off. Proxylang ships three, and you can move between them without redoing the setup.
| Mode | Available on | Your DNS | Search engines |
|---|---|---|---|
| Script tag One line of JavaScript translates the page in the visitor’s browser. Fastest to install and the right choice when you want the language switcher working today. It is not a proxy, and it does not earn you search traffic on its own. | Every plan, including free | Unchanged | See your original language |
| Hybrid Visitors are translated in the browser; search engine crawlers are served a translated page rendered on the server. You get the indexable pages a proxy gives you without moving your DNS. Crawler pages are translated within a per-page token budget. | Start plan and up | Unchanged | Get server-rendered translated HTML |
| Full reverse proxy Every request for a translated URL is rendered on the server for everyone. This is the mode this page is about, and the one that behaves identically for a visitor, a crawler and an AI assistant. | Pro plan and up | Points at Proxylang | Get the same page every visitor gets |
A website translation proxy is a server that sits between your visitors and your website. When somebody requests a translated URL, the proxy fetches the original page from your server, translates the text inside the HTML, and returns the translated page. Your site is never modified, and you do not install anything in it. The translated pages live on your own domain, usually under a language path such as /es/ or /ja-jp/.