Skip to main content

Documentation Index

Fetch the complete documentation index at: https://developers.t2000.ai/llms.txt

Use this file to discover all available pages before exploring further.

You want a trip planned without ten browser tabs: real flight options, how to get around, what the weather will be, and a day-by-day plan. Your agent pulls live Google Flights results, directions between the places you want to see, and the forecast for your dates, then Claude assembles an itinerary. ~$0.08, self-contained — just an origin, a destination, and dates.
This is a durable MPP demo — a sandboxed client can’t fake live flight prices, real routing, or a real forecast. The value is paid, real-time data, not generation.

​
The prompt

Plan a 3-day Tokyo trip leaving from LAX on 2026-07-10, back 2026-07-13.
Use SerpAPI for flight options, Google Maps for getting between Shibuya,
Asakusa, and teamLab Planets, and OpenWeather for the forecast — then give
me a day-by-day itinerary with a flight pick and what to pack.

​
What runs

  1. POST /serpapi/v1/flights — live Google Flights results for the dates (~$0.02)
  2. POST /googlemaps/v1/directions — routes between the places on the list (~$0.02)
  3. POST /openweather/v1/forecast — 5-day / 3-hour forecast for the destination (~$0.02)
  4. POST /anthropic/v1/messages — Claude builds the itinerary + flight pick + packing list (~$0.02)

​
Run it

​
Claude Desktop (MCP)

npm install -g @t2000/cli && t2 init && t2 receive && t2 mcp install
Paste the prompt with any origin, destination, and dates. The agent decides how many routing/forecast passes it needs.

​
SDK

import { T2000 } from '@t2000/sdk';

const agent = await T2000.create();

const [flights, directions, forecast] = await Promise.all([
  agent.pay({
    url: 'https://mpp.t2000.ai/serpapi/v1/flights',
    method: 'POST',
    body: JSON.stringify({
      departure_id: 'LAX',
      arrival_id: 'HND',
      outbound_date: '2026-07-10',
      return_date: '2026-07-13',
    }),
  }),
  agent.pay({
    url: 'https://mpp.t2000.ai/googlemaps/v1/directions',
    method: 'POST',
    body: JSON.stringify({ origin: 'Shibuya, Tokyo', destination: 'teamLab Planets, Tokyo', mode: 'driving' }),
  }),
  agent.pay({
    url: 'https://mpp.t2000.ai/openweather/v1/forecast',
    method: 'POST',
    body: JSON.stringify({ city: 'Tokyo' }),
  }),
]);

const itinerary = await agent.pay({
  url: 'https://mpp.t2000.ai/anthropic/v1/messages',
  method: 'POST',
  headers: { 'anthropic-version': '2023-06-01' },
  body: JSON.stringify({
    model: 'claude-sonnet-4-5',
    max_tokens: 900,
    messages: [{
      role: 'user',
      content:
        `Plan a 3-day Tokyo trip (LAX→HND, 2026-07-10 to 07-13). ` +
        `Pick one flight option with a one-line reason, give a day-by-day itinerary ` +
        `using the routes provided, and a packing list based on the forecast.\n\n` +
        `FLIGHTS: ${JSON.stringify(flights.body)}\n\nROUTES: ${JSON.stringify(directions.body)}\n\nFORECAST: ${JSON.stringify(forecast.body)}`,
    }],
  }),
});

console.log((itinerary.body as { content: { text: string }[] }).content[0].text);

​
Expected output

4 calls · ~$0.08 · ~10s · 0 taps
Flight pick + 3-day itinerary + packing list, grounded in live data

​
Extend it

  • Render the itinerary to a shareable PDF with PDFShift (/pdfshift/v1/convert)
  • Add Google Maps places (/googlemaps/v1/places) to suggest restaurants near each stop
  • Swap SerpAPI flights filters (type, cabin, stops) to compare nonstop vs cheapest
  • Run it for several candidate cities in parallel and let Claude rank them by weather + flight price