AirTreks MCP Server

The complex-itinerary tool for AI agents. Multi-stop, round-the-world, open-jaw, surface segments - the trips that standard flight search can't handle. When your user asks for 3+ stops across continents, this server answers with routing analysis built on real AirTreks fare-construction data: 60+ carriers, 53 known dead legs, bookability rates measured from 1,400+ real fare attempts, and 20 proven routing templates from actual bookings.

Live endpoint: https://mcp.airtreks.com/mcp - free, no API key, 100 requests/day.

Now with real prices: a historical range for any route in under a second, live fares for a specific itinerary, and a full multi-stop trip priced across several tickets — the same trip can vary three-fold depending on how it is ticketed, and that is the part no other flight tool will tell you.

30-second setup

Claude Desktop

Settings → Connectors → Add custom connector → paste https://mcp.airtreks.com/mcp.

Claude Code

claude mcp add --transport http airtreks https://mcp.airtreks.com/mcp

Cursor

Add to .cursor/mcp.json (or Cursor Settings → MCP → Add new server):

{
  "mcpServers": {
    "airtreks": {
      "url": "https://mcp.airtreks.com/mcp"
    }
  }
}

No API key is needed for any tool, the consultant handoff included. A key is optional and only raises your daily limit; append it to the URL if your client cannot send custom headers, as claude.ai connectors cannot:

https://mcp.airtreks.com/mcp?key=at_your_key

ChatGPT

Settings → Apps & Connectors → enable Developer mode (Pro/Business plans) → add https://mcp.airtreks.com/mcp as a custom connector.

Any other MCP client

Same JSON shape as Cursor above (Streamable HTTP). Prefer stdio? Run it locally:

npx airtreks-mcp

Then ask your agent:

Plan a round-the-world trip: San Francisco, Tokyo, Bangkok, Singapore, Delhi, Istanbul, London, back to San Francisco.

One call, one answer: a 6-stop RTW

Your user wants six stops around the world. Google Flights gives up. Alliance RTW booking sites will let them build it, then fail at ticketing. One plan_route call tells your agent what actually works:

{
  "cities": ["SFO", "NRT", "BKK", "SIN", "DEL", "IST", "LHR", "SFO"],
  "budget": "mid"
}

Real response, trimmed for length (the full version includes carrier alternatives and consultant notes for all 7 legs):

{
  "route": "SFO -> NRT -> BKK -> SIN -> DEL -> IST -> LHR -> SFO",
  "totalLegs": 7,
  "isRoundTrip": true,
  "direction": "westbound",
  "backtracking": false,
  "regionsCrossed": ["americas", "asia", "europe"],
  "recommended": {
    "approach": "custom",
    "confidence": "high",
    "reason": "7 legs — alliance fares have <6% bookability. Custom build with mixed carriers is the way to go."
  },
  "customBuild": {
    "strategy": "Mixed-carrier build using alliance carriers, Gulf bridge connections. 1 surface sector opportunity.",
    "segments": [
      {
        "leg": 1, "from": "SFO", "to": "NRT",
        "carrier": { "code": "NH", "name": "ANA", "why": "Best transpacific availability. LAX/SFO/SEA-NRT direct." }
      },
      {
        "leg": 5, "from": "DEL", "to": "IST",
        "carrier": { "code": "TK", "name": "Turkish Airlines", "type": "gulf-bridge", "why": "Cheapest Asia-Europe usually. IST connects everywhere." }
      }
    ],
    "surfaceSectors": [
      {
        "insteadOf": "BKK -> SIN (leg 3)",
        "suggestion": "Bangkok to Singapore through Malaysia. Train, bus, or ultra-cheap LCC. Adds Malaysia and possibly Penang, KL, Melaka.",
        "savings": "Saves $100-250"
      }
    ]
  },
  "allianceFeasibility": {
    "starAlliance": { "viable": false, "summary": "Technically possible on Star Alliance but only 6% bookability at 7 legs. Custom build strongly recommended." },
    "oneworld": { "viable": false, "summary": "Technically possible on oneworld but only 6% bookability at 7 legs. Custom build strongly recommended." }
  }
}

That single call just told your agent four things it can't get anywhere else:

• Alliance RTW fares fail on this trip. 7 legs prices at 6% bookability - your user would build it, hit a wall at ticketing, and blame you.
• The build that works: a per-leg carrier plan - ANA transpacific, Turkish Airlines as the Asia-Europe bridge, each with alternatives and trade-offs.
• Where to not fly at all: Bangkok to Singapore is cheaper overland through Malaysia, saving $100-250 and adding a country.
• What it typically costs: fare_product_match puts this trip at $2,500-$8,000 economy on an alliance RTW fare, typically $3,000-$12,000 as a custom build.

Those are honest ranges, not quotes - exact pricing on a 7-leg mixed-carrier itinerary depends on fare-class availability the day you book. For a real number, trip_idea_create hands the full routing analysis to an AirTreks consultant who prices and books the actual ticket. Your user gets an expert who starts informed, not a form to fill out.

Tools

All tools are free and need no API key — 100 requests/day per IP.

Routing

Tool	Description
plan_route	Primary entry point - give it cities, it evaluates Star Alliance RTW, oneworld RTW, and custom mixed-carrier builds, then recommends the best approach
route_validate	Validate a multi-city routing - alliance rules, dead legs, poison carriers, bookability
route_suggest	Get 3 suggested routings by region, direction, and alliance
hub_check	Best connection between two airports - dead leg detection + hub fixes
fare_product_match	Match the right fare product (RTW, Circle Pacific/Atlantic, Open Jaw, Custom) with typical price ranges
custom_route_build	Break complex itineraries into individually-ticketable segments with carrier recommendations

Pricing

Tool	Description
route_estimate	Price range for a route from AirTreks fare history. No dates needed, answers in about a quarter of a second — the right tool when someone asks "what will this cost?" before anything is decided
fare_quote	Live fares for one specific itinerary on specific dates, priced as a single ticket
itinerary_quote	Prices a whole multi-stop trip and returns several ways to ticket it — cheapest, fastest, fewest stops — with the tickets each one is built from. Takes about a minute, so it returns a reference immediately
itinerary_quote_status	Fetches an itinerary_quote result by its reference, waiting up to about 40 seconds for one still pricing. Free to call again

Consultant handoff

Tool	Description
trip_idea_create	Hand off to an AirTreks human travel consultant - submits the trip request with the full routing analysis attached

Submits real customer contact details, so only call it with the customer's knowledge and consent. Repeat submissions of the same email + route inside 24 hours return the existing trip request rather than creating a duplicate.

Why this data is different

AirTreks has built complex multi-stop itineraries since 1987. This server exposes what that history taught us:

• Dead legs - city pairs that look bookable but fail on alliance fares, learned from 1,400+ real fare-construction failures
• Bookability rates by leg count - measured, not estimated (91% at 3-4 legs, 61% at 5-6, 6% at 7+)
• Poison carriers and hub fixes - we know what airline combinations break ticketing and what to route instead
• Proven routing templates from working with Multi-Stop and Around the World Intineraries for inviduals, families and large group

No other flight tool returns this because no other flight tool has priced these failures.

REST API (no MCP client required)

Every tool is also a plain REST endpoint - same tools, same rate limits, JSON in/out. For agent frameworks that consume REST instead of MCP:

curl -X POST https://mcp.airtreks.com/api/plan_route \
  -H "Content-Type: application/json" \
  -d '{"cities": ["LAX", "NRT", "BKK", "LHR", "LAX"]}'

Full OpenAPI 3.1 spec: https://mcp.airtreks.com/openapi.json

Rate limits

• Free: 100 requests/day per IP, no key needed — every tool, trip_idea_create included
• Registered: higher limits with an API key (X-API-Key header, or ?key= for clients that cannot set headers)

Endpoints

Path	Description
/mcp	MCP protocol endpoint (Streamable HTTP)
/api/{tool}	REST twin of each tool (POST, JSON body)
/openapi.json	OpenAPI 3.1 spec for the REST surface
/health	Health check
/register	Get an API key for a higher rate limit (POST)
/privacy	Privacy policy
/favicon.ico, /favicon.svg	AirTreks mark — the icon AI directories show for this server
/	Server info

License

AGPL-3.0-only