The sky, reviewed.

SetuHQ searches any TESS star's public light curve for transits. A calibrated panel of Sentinels (AI reviewers) and Probes (deterministic detectors) marks the dips, and vetting explains every verdict. Nothing is called a planet until it is one.

Explore the sky See five worked examples ›

Artist's impression

Model from stored values. pi Men c: Rp 1.87 R⊕ (exoplanet.eu, fetched 2026-10-18) over R★ 1.15 R☉ and Teff 5,992 K (TESS Input Catalog v8.2 (MAST)) give Rp/R★ 0.0150; b 0.57 (exoplanet.eu), P 6.268 d. Depth 245 ppm, 3.20 h (quadratic limb darkening, Claret 2017). pi Men's star disc is in its blackbody colour; the planet is an artist's impression (envelope, hot). How the pictures are drawn ›

Worked examples: 5 stars, one pipeline

Each is a stored run on public TESS light curves from MAST, with its data, configuration and code hashed. Every dip is counted by its vetting outcome; a candidate stays unconfirmed, and a rejected dip is not a candidate.

Model from stored values
Star 5,676 K (TIC v8.2) · b Rp/R★ 0.0780 (run fit).

Hot Jupiter

WASP-126

A Jupiter-sized planet on a 3.3-day orbit: deep, easy transits, in the TESS southern continuous viewing zone.

8 dips (8 candidates)

TESS sector 1 · Probes only; Sentinels not run on this example

Case file Plates Run report
Model from stored values
Star 3,532 K (TIC v8.2) · b dashed, Rp/R★ 0.0301 (exoplanet.eu, not recovered by this run) · c Rp/R★ 0.0659 (run fit) · d Rp/R★ 0.0598 (run fit).

Multi-planet system

TOI-270

An M dwarf with three known transiting planets at 3.4, 5.7 and 11.4 days: several periods in one star.

7 dips (5 candidates, 2 flagged)
Found all 5 of c's transits and both of d's; b not found; d seen twice, no period (one transit per 13-day piece of the light curve).

TESS sector 5 · Probes only; Sentinels not run on this example

Case file Plates Run report
Model from stored values
Star 5,175 K (TIC v8.2) · b Rp/R★ 0.0246 (run fit) · c Rp/R★ 0.0238 (run fit).

Small planets near the limit

HD 15337

A bright K dwarf with two small known planets: b (about 1.6 Earth radii, 4.8 days, roughly 350 ppm deep) and c (17.2 days).

3 dips (2 candidates, 1 flagged)
Found 1 of b's 5 transits and both of c's; neither published period was recovered (the BLS period on those dips is not the planet's).

TESS sector 4 · Probes only; Sentinels not run on this example

Case file Plates Run report
Model from stored values
Star 6,422 K (TIC v8.2) · occulter not identified, √depth 0.132 (run fit) · secondary eclipse at SNR 10.9; all 6 dips rejected by vetting.

Known false positive

TOI-236

A TESS Object of Interest the TOI table dispositions as a false positive (FP): a 2.4% dip every 4.5 days.

6 dips (6 rejected)

TESS sector 2 · Probes only; Sentinels not run on this example

Case file Plates Run report
Model from stored values
Star 5,992 K (TIC v8.2) · c Rp/R★ 0.0179 (run fit).

Reference check

pi Men

A sun-like star with a super-Earth on a 6.3-day orbit: the run checked in full against Huang et al. 2018.

7 dips (5 candidates, 1 flagged, 1 rejected)

TESS sector 1 · Probes and Sentinels

Case file Plates Run report

Each stage: the star's disc in its blackbody colour at the Teff its run stores, and what the run measured as a shadow at its fitted Rp/R★ (a planet the run missed is a dashed outline). Sizes are to scale against the disc; where each sits on it is illustrative. No candidate is given planet art. Stages: docs/results/showcase.json; each run's stellar.json (TESS Input Catalog v8.2).

The example stars' colours

Five example stars: 3,532 K to 6,422 K, one pipeline

TOI-270 3,532 K
HD 15337 5,175 K
WASP-126 5,676 K
the Sun 5,772 K
pi Men 5,992 K
TOI-236 6,422 K

Star colour: a Planck spectrum through the CIE 1931 colour-matching functions, in sRGB, normalised to its brightest channel (a hue, not a brightness). Teff: each run's stellar.json (TESS Input Catalog v8.2, MAST); the Sun 5,772 K (IAU 2015 nominal). How the pictures are drawn ›

Planets to scale

The showcase planets, to scale

To scale: one Earth radius is the same width all along the row, so Jupiter is 11.2 times as wide as Earth. Each picture is an artist's impression chosen from the measured radius and the computed irradiation; composition is not measured. Earth, Neptune and Jupiter are the Solar System references (only Earth is drawn with oceans). Radii: The Extrasolar Planets Encyclopaedia (exoplanet.eu), fetched 2026-10-18 (1 RJ = 11.209 R⊕). How the pictures are drawn ›

Real data

How a planet hides in the noise

Real data, cited

3,485 points folded on 6.2680 d: a 292 ppm, 3.07 h dip. Source: runs/tess-pi-men-s1.

How transits are found ›

The panel

Who reviews a dip

Real data, cited

5 of 5 marked it · panel score 1.00, as storedcandidate, unconfirmed

Source: run tess-pi-men-s1, its stored marks and panel score; calibrated weights from docs/results/reviewer_weights_full_panel.json.

Sentinels run on Claude (Anthropic) by default; other providers are supported and labelled. Probes are deterministic matched-filter detectors.

The sky

Every known planet's star, by the year it was found

Real data, cited

8,376 planets shown · transit 4,958 · radial velocity 1,370 · imaging 1,269 · microlensing 386

transit
radial velocity
imaging
microlensing
astrometry
timing
other

Each dot is a catalogued planet at its host star's position (RA and Dec, J2000), coloured by the method that found it, appearing in its discovery year. This still shows 100 of them (every 84th); the full map draws when the page's script runs. SetuHQ did not discover these planets. Source: The Extrasolar Planets Encyclopaedia (exoplanet.eu), fetched 2026-10-18.

Real data, cited

Point at a year, or focus the chart and use the arrow keys.

transit
radial velocity
imaging
microlensing
astrometry
timing
other

Planets per discovery year, stacked by the method that found them. SetuHQ did not discover these planets. Source: The Extrasolar Planets Encyclopaedia (exoplanet.eu), fetched 2026-10-18.

The counts by year and method, as a table ›

Explore the sky ›

Every number links to its proof

The evidence, counted

The reference check: pi Men c, a known planet, found again in run tess-pi-men-s1

5 of 5transits of pi Men c found without being told where to look, each about 8 minutes from its published timing. A known planet, recovered.

2 + 3AI Sentinels and Probes read this run's light curve; each vote counts as much as it scored on planted test transits.

190entries from 21 more stars in the search feed, matching known TOIs. Leads for review, not planets.

1dip still open. The Probes saw it (3 of 3). Both Sentinels passed it by.

We have searched 200 stars for anything that dims; nothing yet stands out above chance. See the evidence ›

What you can do here

What this page does not claim

No confirmed planet. Every dip, signal and feed entry here is a candidate, unconfirmed. Passing vetting is not a confirmation. The match with pi Men c says the planet is already known, not that this run confirmed it.
No planet sizes. No stellar radius is stored, so a depth gives only a planet-to-star radius ratio. The radius-ratio page shows the working.
No people on the panel. The panel is the Sentinels and the Probes: 2 Sentinels and 3 Probes. Sentinels run on Claude (Anthropic) by default; other providers are supported and labelled. Probes are deterministic matched-filter detectors. Their weights rest on calibration with 10 planted and 10 empty light curves each, so they are rough.
Archive values only. This page claims no detection. A TOI or archive match compares a stored period with published values.
No image of the planet. Planet pictures are artist's impressions, chosen from measured radius and irradiation; How the pictures are drawn ›

Learn path

18 steps, from how a survey works to every known exoplanet. Optional: go in order or skip around; any page you open counts. Progress stays in this browser.

Start the path

0 of 18 visited

All 18 steps

Worked examples: 5 stars, one pipeline

WASP-126

TOI-270

HD 15337

TOI-236

pi Men

Five example stars: 3,532 K to 6,422 K, one pipeline

The showcase planets, to scale

How a planet hides in the noise

Who reviews a dip

Every known planet's star, by the year it was found

The evidence, counted

The reference check: pi Men c, a known planet, found again in run tess-pi-men-s1

What you can do here

What this page does not claim

Learn path