Перейти к основному содержимому
Уровень 1

Сайту нужны исправления

Сайт находится в поиске, но не пройдено проверок: 11. Это мешает AI-агентам понять и использовать сайт. Ниже — что сделать в первую очередь.

https://api.starveri.net/
  • 65
    Видимость для AI
    2/5 проверок
  • 29
    Доступность контента
    0/4 проверок
  • 29
    Правила для AI-ботов
    0/3 проверок
  • 0
    Технические протоколы
    0/6 проверок
  • Электронная коммерция
    Не проверено

Что сделать в первую очередь

ВажноОшибка

Markdown content negotiation

Add content negotiation in the route or middleware: render clean page text for Markdown requests while keeping HTML for browsers.

ВажноЧастично

AI bot rules

Add dedicated `User-agent` groups or a wildcard policy with `Allow`/`Disallow`; verify it does not accidentally block required SEO crawlers.

ВажноЧастично

llms.txt

Create a short agent-facing index: a single H1 title, a one-line blockquote summary right after it, `##` sections that list key URLs as `[name](url): note`, prefer Markdown (`.md`) link targets, and serve the file with `Content-Type: text/markdown`.

ВажноЧастично

robots.txt availability

Create or update `/robots.txt`: add rules for generic and AI crawler user agents, include an absolute sitemap URL, and serve it as `text/plain`.

ВажноЧастично

Server-rendered content availability

Enable SSR/SSG or a server-side fallback for public pages; do not leave an empty app root as the only source of content.

Зоны сайта

  • Видимость для AI

    65

    2 пройдено · 2 частично · 1 ошибок

    Можно отложитьЧастично

    Canonical and metadata

    Populate metadata in the layout/template for each public page; do not rely only on client-side head updates. Point each page's canonical at its own clean URL (consolidating tracking/variant query parameters), keep `og:url` consistent with that canonical, add all four required `og:*` properties, and remove `noindex` from pages that should be available to agents.

    СреднеОшибка

    Discovery Link headers

    Add `Link` headers at the web server or application response layer; use absolute URLs and correct `rel`/`type` parameters.

    ВажноЧастично

    robots.txt availability

    Create or update `/robots.txt`: add rules for generic and AI crawler user agents, include an absolute sitemap URL, and serve it as `text/plain`.

  • Доступность контента

    29

    0 пройдено · 2 частично · 2 ошибок

    СреднеОшибка

    llms-full.txt

    Publish the full public content as one Markdown file: an H1 title, a blockquote summary right after it, then `##`/`###` sections of complete text (overview, important pages, usage limits) and no private data. Serve it with an explicit `Content-Type: text/markdown` (or `text/plain`) and `Vary: Accept`, and keep a companion `/llms.txt` index next to it. Do not trim the file just for size — a large valid document passes the check.

    ВажноЧастично

    llms.txt

    Create a short agent-facing index: a single H1 title, a one-line blockquote summary right after it, `##` sections that list key URLs as `[name](url): note`, prefer Markdown (`.md`) link targets, and serve the file with `Content-Type: text/markdown`.

    ВажноОшибка

    Markdown content negotiation

    Add content negotiation in the route or middleware: render clean page text for Markdown requests while keeping HTML for browsers.

    ВажноЧастично

    Server-rendered content availability

    Enable SSR/SSG or a server-side fallback for public pages; do not leave an empty app root as the only source of content.

  • Правила для AI-ботов

    29

    0 пройдено · 1 частично · 2 ошибок

    ВажноЧастично

    AI bot rules

    Add dedicated `User-agent` groups or a wildcard policy with `Allow`/`Disallow`; verify it does not accidentally block required SEO crawlers.

    СреднеОшибка

    Content usage policy signals

    Add consistent policy headers/meta/link rels pointing to usage terms; do not encode legal guarantees in code.

    Можно отложитьОшибка

    Rate-limit hints

    Configure rate-limit middleware/edge rules and return clear headers for 429 or controlled throttling.

  • Технические протоколы

    0

    0 пройдено · 0 частично · 6 ошибок

    СреднеОшибка

    A2A/WebMCP readiness

    Publish agent card/WebMCP JSON only for capabilities that truly exist; add docs links and tests for JSON validation.

    СреднеОшибка

    Agent Skills index

    Create index JSON for public skills, add schema validation, and use stable URLs for skill files or documentation.

    СреднеОшибка

    API Catalog discovery

    Publish API catalog JSON, add links to OpenAPI/JSON Schema documents and a `Link` header; do not include private or internal endpoints.

    СреднеОшибка

    MCP Server Card

    Publish MCP discovery JSON at `/.well-known/mcp` or an MCP Server Card path; when using another URL, add a root `Link` header with `rel="mcp-server-card"`. List only public endpoints and do not expose private tools.

    СреднеОшибка

    OAuth authorization server metadata

    Enable the standard OAuth/OIDC discovery endpoint in the provider or proxy; verify HTTPS URLs, issuer consistency, PKCE `S256`, scopes, and cache headers.

    СреднеОшибка

    OAuth protected resource metadata

    Add well-known JSON for the protected resource, list authorization server URLs, and verify linked server metadata includes PKCE `S256`/scopes without exposing internal resources.

  • Электронная коммерция

    Не проверено