Why
The commuter problems timetable apps don't solve
Most rail apps show you a timetable and hand you the thinking. That works fine for one-off trips. It stops working the moment you're a repeat commuter running on a tight margin.
Why platform uncertainty matters
Platform information at UK stations is often confirmed only a few minutes before departure. For commuters running tight, those few minutes decide whether you make the train or miss it. “Check the board when you arrive” is fine at 10am on a Saturday — it's stress at 08:12 on a Tuesday.
We think platform guidance should be graduated and honest: confirmed when Darwin confirms it, a probabilistic hint when patterns are strong, and explicit uncertainty when it isn't — not a guess dressed up as fact.
Why leave-time decisions are harder than timetable apps suggest
Knowing that the 08:14 leaves at 08:14 isn't the question. The question is: given how long it takes you to walk to the platform from where you are right now, and the realistic buffer at that specific station, should you be walking out the door?
That calculation changes minute by minute. Live delays push it forward. A station with a steep footbridge or a long underpass changes the buffer. Most apps pretend these variables don't exist.
Why alternatives matter, not just alerts
When your usual 08:14 is cancelled, a delay notification doesn't help — it just tells you the problem. You need to know what the next realistic option is, whether it still lands you on time, and whether it's worth jogging.
GetMyTrain surfaces the alternative in the same moment it tells you about the delay. The decision should sit next to the problem.
Why we're launching narrow
We're starting with one corridor — the South West commuter line into Waterloo — because doing one route properly is more useful than doing twenty routes generically. Station-level rules, buffer defaults, and platform patterns are different at every station, and we'd rather get them right in one place first.
Other corridors will follow. Each one will be shaped by what we learn from real commuters on the current corridor.
Our principle is simple: useful now, smarter later.
The AI and behavioural layers are interesting, but they matter only after we've earned trust on the basics. So we're doing the basics properly first.