For AI agents

Give your agent a travel guide with a quality floor

HelpTravel lists only places rated 4.5 stars or higher with at least 1,000 reviews. This page is the machine-readable way in: a Model Context Protocol server, plain JSON endpoints, and a text index. No key, no sign-up, no rate card.

Streamable HTTP MCP 50 curated cities 5 tools

Connect

Point your client at one URL

The server speaks the Model Context Protocol over HTTP. It is public and unauthenticated, so there is nothing to configure beyond the endpoint.

Endpoint

https://helptravel.ai/api/mcp

Claude Code

claude mcp add --transport http helptravel https://helptravel.ai/api/mcp

Any MCP client config

{
  "mcpServers": {
    "helptravel": {
      "type": "http",
      "url": "https://helptravel.ai/api/mcp"
    }
  }
}

Tools

What the agent can ask for

Every response carries the rating and review count behind each place, so an agent can show its work.

list_cities

List every city HelpTravel covers. Curated cities have full editorial guides; discovered cities were generated from a live search and carry places only. Call this first to resolve a city name to a slug.

continent?: Optional filter, e.g. "Europe", "Asia", "North America", "South America", "Middle East", "Africa", "Oceania".

get_city_guide

The places HelpTravel recommends in one city, with the rating and review count behind each. Curated cities also return a short description and an insider tip per place. Every curated place holds 4.5+ stars from 1,000+ reviewers.

city: City name or slug, e.g. "Tokyo" or "new-york".
category?: Limit to one category. Omit for all four.
limit?: Maximum places per category (default 10).

get_city_context

Practical and cultural context for a curated city: when to go, currency, language, what it is known for, neighbourhoods to stay in, dishes to order, and frequently asked questions with answers.

city: City name or slug.

compare_cities

Two curated cities side by side: practical facts, what each is known for, and the top places in each category. Useful when a traveller is choosing between destinations.

city_a: First city name or slug.
city_b: Second city name or slug.

find_places

Search recommended places across every curated city by name, area, or what they are known for. Use it when the traveller describes what they want rather than naming a city.

query: Free text matched against place names, areas, and descriptions.
category?: Limit to one category.
min_rating?: Minimum star rating (default 4.5).
limit?: Maximum results (default 20, max 50).

Try it

One request, no client needed

The endpoint accepts plain JSON-RPC, so curl is enough to see what an agent would get back.

curl

curl -s https://helptravel.ai/api/mcp \
  -H 'content-type: application/json' \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/call",
    "params": {
      "name": "get_city_guide",
      "arguments": { "city": "Tokyo", "category": "restaurants" }
    }
  }'

Plain data

Or skip the protocol entirely

The same content is served as ordinary JSON and text for agents and crawlers that do not speak MCP.

llms.txt

A text index of the site: every city, every guide, one line each.

/api/cities

The full list of covered cities as JSON, curated and search-discovered.

/api/cities/:slug

One city: its places, ratings, review counts, neighbourhoods and FAQs.

Ground rules

What you can do with this

Use it freely, including in commercial products. Attribute HelpTravel and link back to the guide you quote, so a reader can check the ratings for themselves.

Ratings and review counts come from Google Places and were accurate when the guide was written. They move over time. Curated review counts are rounded down to the nearest thousand and marked with a plus, so they are floors rather than exact figures. Place photos come from Wikimedia Commons with their own licences and attributions attached.

The server is a read-only view of published pages. There is no rate limit worth documenting, but it is one small database behind a CDN, so batch your calls rather than crawling every city at once.

Browse the same content as a human at helptravel.ai.