Find the planets the pipelines miss
NASA's TESS telescope has watched some stars for years. A planet on a long orbit crosses its star only once or twice in that time, and the automated searches usually need three crossings. Far Orbits volunteers search those light curves on their own computers, then judge what turns up together.
Search progress
14,275 of 27,458 stars searched
2
Volunteers
528
Slices open
39
Candidates found
22
Vetting votes
How it works
Claim a slice
Grab 25 stars from the shared queue. Every star has years of TESS data.
Run the search
The open-source Python pipeline searches every sector for single dips. About an hour on a laptop.
Results come back
The pipeline uploads by itself. The site cross-checks everyone's results and throws out dips that hit several stars at once (spacecraft glitches).
Vet together
Every surviving signal gets at least three independent reviews, from people and calibrated AI assistants. The best go to professional follow-up.
No telescope or coding needed to vet. The 5-minute guide shows you what to look for.
Use Claude or another AI assistant? Connect it to Far Orbits and it can claim stars, run the search and vet candidates with you.
First pilot result · under review
HD 271568: something crosses this star every 713 days
The pilot run found two identical 1.9% dips two years apart in a Sun-like star that no catalogue lists as special. From those two, we predicted a third crossing in May 2023, and it was there in independent TESS data. The object is about 1.8× Jupiter's size, which fits a giant planet, a brown dwarf or a very small star; only spectra can tell them apart. Either way, it had not been reported.
See all candidates
Strongest candidate · ready for follow-up
TIC 167810285: 23 transits, every 42.48 days, for six and a half years
The search found seven identical 2.8% dips in TESS's 2-minute data. From those we predicted where the others should be, and all 16 turned up in TESS's independent full-sky images from 2018 to 2025. The signal is on the star, perfectly regular, and shows no sign of a second eclipse. The object is about 1.8 times Jupiter's size: a giant planet or brown dwarf on an eccentric orbit, or a small red star. A few spectra will tell which. It is not in any catalogue.
See all candidatesWhat the search can do
- Recovered all 5 transits of the known 225-day planet TOI-4562 b, and its period, blind.
- Catches ~9 in 10 single transits with signal-to-noise above 14 (tested on 480 fake transits in real data). That's roughly Neptune-size and up around a Sun-like star.
- Every known planet candidate (TOI or community CTOI) is removed from the target list up front.
What it can't
- A candidate is not a planet. Most dips turn out to be binary stars, star spots or instrument glitches.
- Confirmation needs telescopes and spectra. The best candidates get submitted to NASA's ExoFOP, credited to the volunteers who found and vetted them.
- Expect roughly one good candidate per several hundred stars. It is a long game.