FortripAI

You supply dates, destination, budget, group size and pace, and the AI itinerary generator returns a day-by-day itinerary with reasons for each choice. It's built for sparse briefs, so "eight days, food-led, mid-budget, late May, no big museums" is enough to begin. Every change you request re-runs the plan instead of patching one line.

The validator reviews an existing plan on flight connections, transit times, daily pacing, routing logic and risk flags such as late arrivals or missing rest days. It works like handing your plan to someone who has been to the city. You get a list of what breaks, with fixes you can accept in the chat.

Flights, trains, buses and ferries are compared between each pair of stops. The tool weighs total travel time, transfers and cost together, which matters on routes where the cheapest leg eats a day. Results feed back into the itinerary, so the schedule updates when you switch modes.

Hotels are matched to the trip rather than sorted by commission. You can compare properties by neighborhood, check cancellation terms, and split a longer stay across two hotels when one area doesn't cover every day. Room-level offers come from live supplier data.

Change a day, swap a city, move a flight, and the plan is checked again automatically. That matters on long planning sessions where a small edit breaks something two days later. The itinerary you export is the one that passed the last check.

FortripAI ships a remote MCP server and a CLI, so agents in Claude Code, Codex or Cursor can search hotels, read live offers, create an order and return a payment link. You confirm and pay yourself, and the agent never charges on its own. The published tools focus on hotels, while itinerary building stays in the web app.