Run it yourself.
SetuHQ is open source. Install it, search a star's light curve with the same panel this site shows, contribute a sweep of the sky, or let your AI assistant use it.
Get started
Install, then search any TESS star's light curve, as each of this site's example runs did.
Install and first run ›Contribute a sweep
Search a slice of a TESS sector on your own machine and send the result, with credit.
How to contribute ›Use it from an assistant
The MCP server lets an assistant run SetuHQ's own code and read its results.
Connect over MCP ›Data and files
Every catalogue on this site as a download, and the reports behind the evidence.
Downloads ›Install and first run
SetuHQ needs Python 3.11 or newer.
pip install "setuhq[tess] @ git+https://github.com/ashfordeOU/SetuHQ"
setuhq run configs/showcase/tess-wasp-126-s1.toml --out runs/tess-wasp-126-s1 # or any star: copy a config, change its TIC and sectors
setuhq crossmatch runs/tess-wasp-126-s1 # the NASA Exoplanet Archive's planets for the star
setuhq site runs/tess-wasp-126-s1 --out site # a website from your run (several runs: list them all)Every example on this site has its configuration in configs/ and its one-line command in docs/SHOWCASE.md. The run downloads the light curve from MAST, cleans it, lets the panel mark dips, vets each signal and writes every result with its provenance. The Probes run with no account; the Sentinels need an API key for their provider. Sentinels run on Claude (Anthropic) by default; other providers are supported and labelled. Probes are deterministic matched-filter detectors.
Contribute a sweep
A sweep searches every 2-minute target of a TESS sector and keeps every dip it finds, classified, with nothing thrown away. You can run a slice and send it in:
setuhq sweep configs/sweep.toml --sector 5 --limit 200 --workers 3
setuhq bundle my-sweep/sector_0005.json.gz --contributor "Your Name" --out bundle/Open a pull request with the bundle. A maintainer checks it and merges it into Everything that dims, with your name on every entry it brings. See CONTRIBUTING.md.
Use it from an AI assistant (MCP)
pip install "setuhq[tess,mcp] @ git+https://github.com/ashfordeOU/SetuHQ"
setuhq-mcp --list-toolsSet-up for each assistant is in docs/MCP-INSTALL.md.
Data and files
- The catalogue of known planets (JSON) and every planet on the Galaxy map (CSV), from exoplanet.eu.
- Everything that dims (JSON): every entry of SetuHQ's sweep, with its credit.
- WASP-126 (Hot Jupiter): the run report, the plates and the case file of run tess-wasp-126-s1.
- TOI-270 (Multi-planet system): the run report, the plates and the case file of run tess-toi-270-s5.
- HD 15337 (Small planets near the limit): the run report, the plates and the case file of run tess-hd-15337-s4.
- TOI-236 (Known false positive): the run report, the plates and the case file of run tess-toi-236-s2.
- pi Men (Reference check): the run report, the plates and the case file of run tess-pi-men-s1.
- The benchmark, the leaderboard and the paper: docs/ in the repository.
SetuHQ's own data, benchmark, leaderboard and paper are under CC BY 4.0; the code is under the AGPL-3.0 (a commercial licence is available). Cite it with CITATION.cff.
Technical details: where the figures come from
The candidate pages show figures from the stored runs (tess-wasp-126-s1, tess-toi-270-s5, tess-hd-15337-s4, tess-toi-236-s2, tess-pi-men-s1), made with the current code.
Every number on the candidate and evidence pages is read from stored files: the run directory, the search catalogue and the calibrated reviewer weights. Runs and configuration lists the files, the configuration and its hash.