Chapter Nineteen · Live feed
Live globe · every tracked object, drawn where it is
Renderquerying…
Dimmed on the globe = in Earth’s shadow.
Shells classify Starlink by perigee altitude. Boundaries derived from live TLEs on 2026‑08‑23 — the constellation clusters at 440‑500 km (primary). Deorbiting Starlinks draw as decaying regardless of shell.
A cinematic view of near-Earth space — satellites drawn at readable size against a real-time Earth. Hover any object to identify it.
Observation is the default cockpit-style view — satellites at readable size, gentle twinkle, real time. Switch to Starfield for a schematic showing every one of the ~12,500 tracked objects at once (Starlink defaults hidden either way; toggle it on in the class matrix above to see the full LEO population). Detailed models draw the nearest satellites as textured 3D craft.
Surface imagery is NASA’s Blue Marble: Next Generation with topography and bathymetry — the August 2004 MODIS composite, courtesy NASA Earth Observatory and Reto Stöckli (NASA/GSFC). It is a cloud-free true-colour map of the land surface, not a live view: the coastlines are real, the season is August, the weather is nobody’s. Positions are propagated in your browser from elements that refresh daily — the motion is real, the element set is a day old at most. Your marker appears once you have given the panel above a position; it is passed between sections inside this page and still never leaves your device. The lit hemisphere is computed, not painted — it moves with the real terminator. The spacecraft are drawn roughly a hundred times oversize or they would be smaller than a pixel: their positions and attitudes are real, their sizes are not. Colour is the object’s kind, which is read from its catalogue name rather than declared by the feed — treat it as a reading, not a classification.
Low Earth orbit is screened every day for objects on converging paths. Most of what it finds passes quietly and is never recorded anywhere — the forecast rolls forward, and yesterday's near-misses disappear. This is the ledger of what the screening found, and where the lowest objects are right now.
Live positions · propagated to now
Positions are computed in your browser from the most recent orbital elements (SGP4), advancing each second. The elements refresh on the daily feed; the positions are live. Objects with 6-digit catalogue IDs show measured perigee until full live support lands.
Ranked by maximum collision probability. The bar plots miss distance on a logarithmic scale from 1 m to the 5 km screening threshold.
as published · static until the feed loads
| Primary | Secondary | Separation (log) | Max Pc | Rel speed km/s | Closest approach |
|---|---|---|---|---|---|
| STARLINK-30464 | STARLINK-35568 | 26 m | 0.1482 | 1.935 | 2026 Aug 27 · 00:24:49 UTC FORECAST |
| STARLINK-30553 | STARLINK-36554 | 50 m | 0.1309 | 6.163 | 2026 Aug 23 · 21:47:26 UTC NEAR |
| STARLINK-5756 | STARLINK-36799 | 48 m | 0.1256 | 10.302 | 2026 Aug 25 · 14:59:36 UTC FORECAST |
| STARLINK-30714 | GUOWANG TEST OBJECT C | 30 m | 0.1021 | 10.509 | 2026 Aug 23 · 19:35:22 UTC NEAR |
| QIANFAN-199 | UNKNOWN | 5 m | 0.0911 | 12.384 | 2026 Aug 21 · 10:55:56 UTC IMMINENT |
| BRO-12 | STARLINK-33983 | 25 m | 0.0761 | 14.909 | 2026 Aug 26 · 06:52:04 UTC FORECAST |
| STARLINK-11617 | STARLINK-38052 | 64 m | 0.0738 | 10.949 | 2026 Aug 25 · 13:02:07 UTC FORECAST |
| STARLINK-30069 | STARLINK-33831 | 72 m | 0.0621 | 9.845 | 2026 Aug 23 · 17:27:55 UTC NEAR |
| STARLINK-37363 | PEGASUS R/B | 44 m | 0.0450 | 12.313 | 2026 Aug 21 · 11:21:40 UTC IMMINENT |
| STARLINK-2399 | STARLINK-5043 | 94 m | 0.0416 | 7.428 | 2026 Aug 27 · 13:41:45 UTC FORECAST |
The same window, ranked by absolute miss distance instead of probability. It is a different list — and that difference is the story.
as published · static until the feed loads
| Primary | Secondary | Separation (log) | Max Pc | Rel speed km/s | Closest approach |
|---|---|---|---|---|---|
| QIANFAN-199 | UNKNOWN | 5 m | 0.0911 | 12.384 | 2026 Aug 21 · 10:55:56 UTC IMMINENT |
| PRC TEST SPACECRAFT 4 | CZ-2F DEB | 15 m | 0.0192 | 0.000 | 2026 Aug 20 · 22:57:32 UTC IMMINENT |
| ONEWEB-0028 | ARIANE 5 R/B | 16 m | 0.0384 | 11.752 | 2026 Aug 25 · 05:14:49 UTC FORECAST |
| PRC TEST SPACECRAFT 4 | CZ-2F DEB | 17 m | 0.0192 | 0.000 | 2026 Aug 21 · 00:33:50 UTC IMMINENT |
| PRC TEST SPACECRAFT 4 | CZ-2F DEB | 18 m | 0.0192 | 0.000 | 2026 Aug 20 · 22:12:00 UTC IMMINENT |
| PRC TEST SPACECRAFT 4 | CZ-2F DEB | 19 m | 0.0192 | 0.000 | 2026 Aug 21 · 02:10:08 UTC IMMINENT |
| OPS 5712 (P/L 153) | THORAD AGENA D DEB | 20 m | 0.0045 | 14.193 | 2026 Aug 21 · 08:59:21 UTC IMMINENT |
| PRC TEST SPACECRAFT 4 | CZ-2F DEB | 20 m | 0.0192 | 0.000 | 2026 Aug 20 · 23:48:36 UTC IMMINENT |
| PRC TEST SPACECRAFT 4 | CZ-2F DEB | 21 m | 0.0192 | 0.000 | 2026 Aug 21 · 03:46:26 UTC IMMINENT |
| PRC TEST SPACECRAFT 4 | CZ-2F DEB | 22 m | 0.0192 | 0.000 | 2026 Aug 21 · 01:25:13 UTC IMMINENT |
The two lists rarely agree. Probability rises when two tracked, active satellites converge with tight uncertainty — which increasingly means one constellation meeting itself. Raw proximity is dominated by spent rocket bodies and fragments, which pass closer but carry looser position estimates. One list describes crowding. The other describes debris.
Appearances across both lists, by operator.
as published · static until the feed loads
Counted across 20 listed encounters, of which 6 were between two satellites of the same fleet.
Catalogued objects whose orbits are decaying, ordered by perigee — the lowest point of the orbit, and the best available proxy for how little time is left.
as published · static until the feed loads
| Object | Perigee km | Semi-major axis km | Inclination° | State |
|---|---|---|---|---|
| STARLINK-5190 | 172.851 | 6557.796 | 53.15 | IMMINENT |
| STARLINK-1623 | 173.768 | 6557.094 | 53.01 | IMMINENT |
| DONGPO 05 | 188.927 | 6574.955 | 97.54 | IMMINENT |
| TRISAT-2 (RUVDSSAT1) | 193.335 | 6582.898 | 97.35 | IMMINENT |
| STARLINK-1830 | 197.165 | 6580.763 | 53.02 | IMMINENT |
| YINLI-1 R/B | 201.565 | 6601.880 | 97.77 | TRACKED |
| STARLINK-4437 | 205.852 | 6589.738 | 97.59 | TRACKED |
| STARLINK-1695 | 208.804 | 6590.090 | 53.02 | TRACKED |
| STARLINK-1718 | 213.633 | 6596.691 | 53.03 | TRACKED |
| STARLINK-6137 | 216.722 | 6599.694 | 43.00 | TRACKED |
| LEMUR-2-SPECIAL-K | 218.887 | 6606.533 | 97.55 | TRACKED |
| STARLINK-2041 | 220.061 | 6606.142 | 53.14 | TRACKED |
| STARLINK-1597 | 222.916 | 6602.865 | 53.02 | TRACKED |
| OBJECT DB | 225.406 | 6610.164 | 97.35 | TRACKED |
| STARLINK-5533 | 227.767 | 6609.766 | 97.52 | TRACKED |
| OBJECT A | 230.381 | 6617.265 | 97.33 | TRACKED |
| QIANKUN-1 | 232.217 | 6636.291 | 96.13 | TRACKED |
| STARLINK-1626 | 233.748 | 6613.584 | 53.02 | TRACKED |
| PAN SAT (PO-34) | 234.170 | 6614.101 | 28.43 | TRACKED |
| STARLINK-4310 | 236.223 | 6614.794 | 97.58 | TRACKED |
This list does not predict re-entry dates. The upstream catalogue publishes orbital elements, not re-entry forecasts, and a credible forecast needs atmospheric density modelling this page does not do. Perigee altitude is shown because it is measured, not estimated.
The same element sets, resolved against one place on the ground. Give the page a position and it computes when each tracked object crosses your sky, which way it rises and sets, how high it gets, and whether it will be lit by the Sun while your sky is dark.
Loading the element set…
| Object | Rises | Track | Highest | Duration | Naked eye |
|---|
Your position is used in this browser and nowhere else. It is not sent to the feed, not written to storage, not put in the address bar, and not logged. Reloading the page forgets it. Every figure in this table is computed here on your device from the published elements — nothing about where you are crosses the network.
The tracked set is stations, the brightest-objects catalogue and decaying orbits — not the whole sky. “Naked eye” means the object is above the horizon floor, sunlit, and your sky is past civil twilight — sampled at three points of the pass. It is a geometric verdict, not a brightness prediction: a lit object can still be too faint to see.
Each daily build stores the SHA-256 of the exact upstream payload it read, so any figure on this page can be traced back to the bytes it came from.
Not for operational use. These are screening results, not conjunction assessments. Collision probabilities depend on a covariance model and should not be used for manoeuvre decisions. Operators fly on data from their own tracking providers.