How Trypn uses AI — and where it doesn’t
AI recommends. Sources prove.
Beta notice — pending professional review. This page describes how Trypn behaves today. Nothing here is a guarantee of accuracy, availability or fitness for any particular trip.
What AI can help with
Turn what you say into planning ideas, put things in a sensible order, read the notes and screenshots you save, and explain a choice in plain words. It works only with material that already exists: your words, your saved items, and places that came from a source.
What must come from a source
Introduce a real place, or state a fact about one — the name, where it is, when it opens. Those come from a place data provider or from something you added yourself, and each one carries where it came from and when it was last confirmed. An idea without a source stays marked as an idea and never becomes a place in your plan.
Freshness and staleness
Provider information goes out of date. Trypn shows when a detail was last confirmed and says plainly when it could not confirm something, rather than filling the gap. A detail that has gone stale is labelled stale — it is not quietly refreshed with a guess.
Your reports versus provider facts
If you tell Trypn something is wrong or closed, that is recorded as your report and shown as your report. It does not overwrite the provider's record and it is not presented to anyone as verified. A person reviews those separately.
Always check these yourself
Opening times, prices, tickets and closures before you rely on them. And anything with consequences — visas, entry rules, health requirements, insurance, safety and legal questions. Trypn does not give visa, health, legal or safety advice, and should never be your only source for them.
When something is switched off
Assisted steps can be paused during the beta. When that happens Trypn says so and falls back to plain, non-assisted behaviour. Your plan stays readable and editable either way.