# Website Directory / 导航站

> One verified destination per registrable root domain / 每个可注册根域名只展示一个目标站.

## Public access / 公开读取

Public pages render readable HTML. Chinese and English interfaces are /zh and /en. Search and read the published catalog through the same-origin GET endpoint:
公开页面输出可读取的HTML；中文和英文入口分别为 /zh 和 /en。以下GET接口只返回公开内容。

- Search: http://60.205.176.97/navigation/api/agent?locale=en&q=example&page=1
- Read: http://60.205.176.97/navigation/api/agent?locale=en&id=RESOURCE_ID&page=1
- Resource identifier: 32-character website id, returned in search results.
- Follow nextPage or nextCursor when provided. Text segments are bounded; the returned source URL links to the full page.
- Search results may be empty. Unavailable content returns 404; service outages return 503. Do not invent results.

## Browser tools / 浏览器工具

- navigation_search: search the same public catalog, with bounded queries and result pages.
- navigation_read: read a resource returned by search, with paginated text when needed.

Listings are bilingual metadata summaries with verified root-domain identity. Aggregator websites, merged aliases and unverified destinations are absent from public results. A website description does not guarantee the availability, quality or terms of the external service.
目录简介来自目标根站元信息并提供中英版本；公开结果排除聚合站、合并别名与待核实站点。简介不保证外站可用性、质量或条款。

## Compatibility / 兼容性

WebMCP is a progressive enhancement based on the current document.modelContext draft. Native calls require a secure context and a browser implementation enabled by its trial or testing support. HTTP deployments still provide HTML, this Markdown guide and public GET resources. No polyfill advertises unsupported native tools.
WebMCP按当前document.modelContext草案渐进接入，需要安全上下文及已启用支持的浏览器。HTTP部署仍可读取HTML、Markdown指南与公开GET资源；不伪装浏览器原生支持。

- [Site overview / 网站索引](http://60.205.176.97/navigation/llms.txt)