Terminal1
Turn airport time into business opportunities. Meet another professional for 30 minutes during your layover, matched by AI.
Why
In late 2024, I was eating dinner alone during a three-hour layover in Munich. I’d taken around 60 flights that year. Airports kept putting me near thousands of people, yet I rarely met anyone. Somewhere nearby might be a future collaborator, customer or friend, with the same time to spare.
That became the idea for Terminal1: arrange a useful 30-minute conversation while two people’s journeys overlap. Matching needed to consider both what they wanted from a meeting and the practicalities of being in the same airport at the same time.
It was also a chance to explore a question I was asking about AI agents in 2024–2025: what would a service look like if agents handled the operations from the beginning? Terminal1 gave that question a concrete job, from reading a flight ticket to proposing a meeting.
How
The experience was built around Maya, an assistant you could email at maya@terminal1.app:
- Create a profile with LinkedIn, then send a CV to give the matching system more context.
- Forward your flight details so Maya can work with your airport and travel times.
- Explain who you’d like to meet: a potential partner, investor, employer or someone with shared interests.
- Review a proposed match for a 30-minute meeting that fits both travelers’ schedules.
Behind Maya, I organized the agents like an office from the 1950s. Reception read and routed the mail. Specialists processed CVs and tickets. The matching agent evaluated compatible travelers, and a communication agent wrote back. A simple web dashboard made the information visible too.
A useful match had to work in both directions: an investor looking for an AI company and a founder seeking investment, for example. Sharing an airport was only the first filter.
I started the project after that layover and published the agent architecture in May 2025. Development was later discontinued, and the code is open source. Terminal1 remains a worked example of an agent-operated service with email as its front door.