Is this flight price actually good?

The deal oracle your AI agent can call — verdicts backed by price history collected 24/7 on nomad corridors. One URL, no API key, no signup.

Connect your agent

One URL — no keys, no signup. Tools appear in your agent automatically.

MCP endpoint
Claude Code
Claude Desktop
Hermes

Setup guides & example agent missions: github.com/corteslabsxyz/Flait-MCP ↗

READ · FREE is_good_deal — “is $470 LAX→BKK actually good?” → verdict vs real history
find_deals — best current fares across nomad corridors
route_history — the raw daily price series
ACT · GUARDED set_watch — standing watch on ANY route (watching starts price collection on it); optionally armed to auto-book under the user's spend cap
watch_status — poll a watch and its bookings with the key set_watch returns
Writes are guarded. Booking never fires silently: the human gets a “booking in 30 minutes” email with a stop button, spend is capped at their number, and one watch books at most one flight.

Lifecycle webhooks — signed callbacks to your agent:

hitbooking_scheduled bookedbooking_failed booking_stopped
?

“Is this fare good?”

Your agent calls the oracle before anyone books a dud. One GET, real answer.

👁

“Watch Bali under $500.”

Agent sets the intent; Flait carries it 24/7 so your agent doesn't have to run.

⚡

“Book it when it hits.”

Mandate with a cap. Warn-then-book. Confirmation code back to your user.

Free to ask. Paid to act.

Reads stay free — verdicts, deals, history, for every agent. A small fee ($9 + 4% of fare, max $49) rides on completed bookings only. Beta: everything free, booking in test mode.