How it works
A transit is a line on the ground, not an event in the sky
The useful question is never "when is the next transit". It is "which one, and exactly where do I stand". This page explains how that second question is answered — the geometry, the scoring, the alerting, and how far any of it can be trusted.
The geometry
Why five kilometres matters
The spacecraft appears at the centre of the disk only for observers standing where the line from the target's centre, through the spacecraft, meets the Earth. Trace that line over time and you have the centre of the corridor exactly. Either side of it there is a few kilometres of ground from which some part of the crossing is still visible. Beyond that there is nothing to photograph — the spacecraft simply passes near the disk instead of across it, and the trip was wasted.
Four steps
From "where am I" to "this one, at this spot"
Say where you are, and how far you will go
An approximate location and a travel budget — from "where I am standing" to a few hundred kilometres. Nothing else is needed to begin.
It finds and ranks every transit
The search runs on your own device and returns the crossings of the Sun and Moon within reach, ordered by how good a photograph each one is actually likely to make.
It tells you whether it is worth the trip
A verdict with its reasons, an estimated drive, a departure time, and the exact point to stand — shown on a map with the corridor drawn around it.
It keeps watching after you close it
Set a standing rule and it checks on your behalf, alerts you when something good appears, and tells you again if the corridor moves after you have been told where to go.
A ranked answer, not a table
A score you can argue with
Every transit is scored on nine weighted factors, and each one is reported on its own. So when an event ranks below another, you can see precisely why — a grazing chord, a target too low, an hour that costs more than the photograph is worth.
| Factor | What it measures | Weight |
|---|---|---|
| Centrality | How close to the middle of the disk the crossing runs | 20 |
| Target altitude | How high the Sun or Moon sits above the horizon | 15 |
| Travel | Distance, judged against the budget you set | 13 |
| Apparent size | How large the spacecraft appears from where you stand | 11 |
| Prediction reliability | How much the orbital data could still move | 10 |
| Sky | Forecast cloud cover at the observation point | 10 |
| Duration | Seconds the spacecraft spends on the disk | 9 |
| Corridor width | How much room there is for error either side | 7 |
| Hour | What the local time of day costs you | 5 |
Travel is judged against the budget you set rather than against an arbitrary distance, because ten kilometres is a long way to someone who chose "where I am standing" and nothing at all to someone who chose two hundred. Weather is one factor among the nine and never touches the astronomy. When no forecast is available its weight is redistributed rather than assumed clear — an unknown sky must not flatter an event.
Watching
Standing rules
The good ones arrive a few times a year, and nobody catches those by chance. So you set the rule once — this radius, this quality, this much warning — and the checking happens without you. Everything it queues appears in the app whether or not a notification was ever delivered.
A transit worth your attention
Something above the quality threshold you set has appeared within your radius.
Reminder before the event
At the lead time you chose — long enough to pack, drive and set up.
The corridor has moved
Fresh orbital data shifted the centre line for an event you were already told about.
Do not drive — it is off
An event you were told about is no longer reachable, or no longer happens at all.
The forecast has worsened
Cloud cover for a saved event has changed materially. Off unless you switch it on.
And nothing else
The complete set. Notification permission is requested only once a rule exists, never on first launch.
Alerts arrive at night, and acting on one means getting dressed and driving somewhere. That interruption is expensive enough that the threshold is set high and the volume kept deliberately low. An alert that turns out not to have been worth a night's sleep is how an alerting product gets muted.
Measured, not asserted
How far the predictions can be trusted
419 events were collected from an established reference tool and compared one at a time. The propagator was separately measured against NASA's own published ISS ephemerides.
What we do not claim. We do not claim to be more accurate than the established transit finders. We claim to agree with them closely, to have measured that agreement rather than assumed it, and to tell you honestly how uncertain any individual prediction is. That comparison found three real defects in our own calculations, and all three were fixed before any of this was said out loud.
A prediction can still move. Orbital data is reissued several times a day, no terrain or buildings are modelled, and a forecast is a forecast. Every event carries its reliability grade and the sentence explaining it — and the honest advice is always to re-check on the day.
Questions
Frequently asked
What exactly is a transit?
It is the moment a spacecraft passes directly between you and the Sun or the Moon, appearing as a small dark silhouette crossing the bright disk. It is not the same as watching the ISS drift across the night sky: a transit lasts about a second and is visible only from a narrow corridor on the ground, which is why it has to be planned rather than waited for.
Which spacecraft can it find transits for?
The International Space Station, the Chinese space station Tiangong, and the Hubble Space Telescope. The catalogue is data rather than code, so adding an object is a data change — but objects too small to resolve as a silhouette are deliberately left out rather than added to lengthen a list.
How accurate is it?
We do not claim to be more accurate than the established transit finders. We claim to agree with them closely, and to have measured that agreement instead of assuming it: 419 events compared one by one, a median difference of 0.137 seconds in predicted time and 85 metres in centre line within 10 kilometres, plus a median error of 0.158 kilometres against NASA's own published ISS ephemerides. That comparison found three real defects in our calculations, all of which were fixed.
Can a prediction still be wrong?
Yes, and the app says so rather than implying otherwise. Orbital data is reissued several times a day and the corridor moves with it, particularly after a station reboost. Every event carries a reliability grade and the reason behind it, weather forecasts are probabilistic and only ever affect ranking, and no terrain, buildings or trees are modelled. Always re-check on the day.
Does it need my precise location?
No. The search runs on your device, so your coordinates are not sent anywhere in order to perform it, and a place you save is stored to roughly two and a half kilometres — deliberately coarser than a transit corridor. The Android app does not request fine location at all.
Do I need an account?
Not to start. The first session is anonymous, and you can save a place and set a standing alert without giving an email address. Signing in with an email is optional, and uses a single-use link rather than a password.
What equipment do I need?
A camera capable of a fast burst, and a long lens or a telescope. The app works out the field of view and how many pixels the spacecraft will span for the sensor and focal length you tell it about, so you can see what your own kit will resolve.
Go and photograph something
Free to use, with monitoring and alerts included. Plus extends the search window, the travel radius, and the number of places you can watch.