May 13, 2026 · 7 min read · Use cases

Corrected 14 November 2026: the earlier version showed a name search on /v1/cities and context fields the API does not have (a list of airports, a sentiment trend). The calls and fields below are the real ones.

Travel API for B2B booking engines: a hotel inventory companion.

Booking engines have a clean view of hotel inventory and a near-empty view of everything around it. Pair your inventory with a destination metadata API and the same product starts answering questions your travelers actually ask.

The asymmetry

The typical B2B booking engine ships with rate, availability, and room photos. It does not ship with: which UNESCO sites are within 50 km of the property, which tourism circuit the city is on, which airports serve it, or how the destination is currently trending in sentiment.

This is the gap. The booking engine knows the hotel. It does not know the destination. Travelers ask about the destination first.

Autocomplete: master_score ordering

Start with the search box. Most engines order autocomplete alphabetically or by popularity-as-bookings. That biases hard toward the same five cities. /v1/cities returns a master_score per city that ranks destinations by genuine travel relevance, so the top of the list is what users actually mean.

GET /v1/cities?iso_alpha2=IN&min_score=60&limit=200

[
  ...,
  { "id": "714bddd6-7ffa-4081-9449-0bff4f521451", "name": "Agra", "master_score": "69.38", "tier": "Top", ... },
  ...
]

/v1/cities has no name search: load the ranked list for your markets once, match what the user types against it in your own search component, and order the matches by master_score. To resolve one typed name, the find_city tool (POST /v1/agent/tool/find_city) takes the name and the country code.

Destination detail panels

When a user picks a city, fire one call to /v1/cities/{id}/context. It rolls up everything you need for a destination panel into a single payload:

GET /v1/cities/714bddd6-7ffa-4081-9449-0bff4f521451/context

{
  "city_id": "714bddd6-7ffa-4081-9449-0bff4f521451",
  "name": "Agra",
  "iso_alpha2": "IN",
  "admin": { "state_code": "IN-UP", "state_name": "Uttar Pradesh", "district_name": "Agra district", ... },
  "airport": null,
  "tourism_circuits": [{ "slug": "golden-triangle", "name": "Golden Triangle", "sequence_no": 2, ... }],
  "heritage_monuments_nearby": [
    { "name": "Agra Fort", "category": "World Heritage Site", "distance_km": 0.62, ... },
    { "name": "Agra Fort", "category": "Monument of National Importance", "distance_km": 0.62, ... }
  ],
  "sentiment": [],
  ...
}

The heritage list holds the ASI monument records attached to the city itself. airport is null where we hold no nearest airport for the city, and sentiment is filled on paid plans only.

Render that as a side panel next to your hotel results. The booking flow stays unchanged. The destination context is now explicit instead of left to the user's imagination.

Worked integration

async function destinationPanel(cityId) {
  const ctx = await tmai(`/v1/cities/${cityId}/context`);

  return {
    headline: ctx.name + ", " + ctx.admin.state_name,
    heritageCount: ctx.heritage_monuments_nearby.length,
    nearestAirport: ctx.airport,          // may be null
    onCircuit: ctx.tourism_circuits[0]?.slug ?? null,
  };
}

Why this is a moat for the booking engine, not the data vendor

The metadata is commodity. The integration into your booking flow is not. Once your search ranks by master_score and your detail page renders heritage and circuits, your engine looks like a vertical product. Competitors using a generic geo-API still show "Population: 1,033,756" and call it a destination panel.

Pricing implication

A booking engine doing 100k searches/month with 5% click-through to a detail page is roughly 105k API calls. That fits comfortably in the Growth tier at $249/mo — well under any reasonable revenue line on inventory bookings.

Sign up — free See India coverage →