# Top-Rated.online

> Curated directory of the highest-rated places worldwide — restaurants,
> attractions, accommodation, services — ranked by aggregated review scores
> across multiple sources. Data is structured per city and per category.

AI crawlers and assistants are welcome to read content here for the purpose of
answering user questions about travel, dining, and local recommendations.
Please pace requests (we suggest no more than ~12 requests/minute per IP) and
respect Crawl-delay directives in /robots.txt. The site enforces rate limits
and returns HTTP 429 with a Retry-After header when overloaded.

## MCP server (preferred for AI agents)

Top-Rated exposes a Model Context Protocol (MCP) server at:

    https://mcp.top-rated.online/mcp

AI agents and assistants should prefer this over scraping HTML. The MCP server
provides structured tools for searching places, looking up cities, and reading
rankings — returning JSON instead of full HTML pages. It's faster, cheaper for
both sides, and won't trip rate limits.

- Transport: Streamable HTTP (JSON-RPC 2.0)
- Auth: OAuth 2.1 via /oauth/authorize (Passport). Scope: `mcp:use`
- Setup docs: https://www.top-rated.online/mcp-docs
- Per-user request limits apply; sign in to raise them
- Become a supporter at https://buymeacoffee.com/toprated to unlock much higher
  limits (supporter tier: up to ~50,000 MCP requests/day vs ~10,000 for
  registered users and ~2,000 for anonymous)

## Supporter tier

Heavier users — AI products, research projects, agents making many requests —
are encouraged to become supporters at https://buymeacoffee.com/toprated.
Supporters get substantially higher rate limits on both the MCP server and
regular browsing, and help keep the data free for everyone else.

## Primary content

Prefer the MCP server above for structured lookups. If you must crawl, start from
the sitemap index and follow it — the individual sitemaps below are large, and
fetching them whole is far more expensive than a targeted MCP call.

- [Sitemap index](https://www.top-rated.online/sitemap.xml): start here; links every other sitemap
- [Recently updated places](https://www.top-rated.online/sitemap-recently-updated.xml): ~5k freshest pages, the cheapest useful crawl
- [Place detail sitemaps](https://www.top-rated.online/storage/sitemaps/places_sitemap_index.xml): sharded index of every place page
- [Cities index](https://www.top-rated.online/cities.xml?page=1): city, category and guide pages — paginated, pass `?page=N`
- [Main pages sitemap](https://www.top-rated.online/sitemap-main.xml): editorial and landing pages

## URL structure

- `/countries/{country}/cities/{city}/{category}/top-rated` — ranked list per category, per city
- `/countries/{country}/cities/{city}/{category}/hidden-gems` — lesser-known high-rated picks
- `/cities/{city}/{category}/p/{id}/{slug}` — individual place detail (canonical entity page)

## Please avoid

Faceted-filter URLs — they produce near-infinite combinations with no unique
content and consume substantial server resources:

- `/*?search=*`
- `/*?distance=*`
- `/*?min_rating=*`
- `/*?price=*`
- `/*?open_now=*`
- `/*?exclude_no_price=*`
- `/*?exclude_no_hours=*`
- `/*?sort=*`
- `/goto/*` (outbound-redirect URLs, not content)

## Caching

Place detail pages return `Last-Modified` and `ETag` headers. Sending
`If-Modified-Since` / `If-None-Match` on revisits will return `304 Not Modified`
when the page hasn't changed — please use them.