TESS observation laboratory

Search and verification of unusual changes in star brightness.

Connect device
In queue…
Calculations completed…
Devices in operation…
Confirmed discoveries0

Light curve

Method diagram
Local brightness increase diagramMethod illustration, not TESS measurement and not detected object.TimeRelative fluxCheck regionStartEnd of section
Local brightness increase requires verification of original photometry and instrumental interference.

Computation status

Updating...

Getting main node status...

Getting participant status...

Loading statistics.

DataTESS public observations
SelectionBrightness changes
VerificationPhotometry and systematics
ResultReport for review

Participation in calculations

Personal computer or VPS
  1. Get code

    One device, activation period 72 hours.

  2. Pass to agent

    They will verify the source code and agreed launch location.

  3. Check first report

    Test calculation and file receipt confirmation.

Connect device

Code allows only taking tasks and returning results. It does not provide access to other computers.

Save personal note for access management.

Method and conditions

Load and management

The main calculation runs on a processor with a single computational thread. GPU not required. To start you need Python 3.11-3.13, 2 GB free space and internet. Initial installation of scientific libraries may download several hundred megabytes beyond the observation limit.

No autostart. Close the session with Ctrl+C or file 03-STOP.cmd. Report storage limit: 500 MB. You can increase runtime as you wish.

What data the server receives

Random participant number, key hash, job status, numerical results, graph, HTML report and scientific calculation log from open TESS observations. Also saved are your specified name, optional Zooniverse nickname, device name, OS, selected location, program version, session state and stop reason. This information is available to you and the lab owner, but is not published in general statistics. Passwords, email and contents of personal folders are not requested. Device key is stored locally. To protect code issuance, IP hash and hourly window are temporarily stored and deleted after window expiration on new requests.

Personal codes and keys are not recorded in the request log for this page. The browser stores the note locally so you can return and revoke access. The page has no advertising trackers. Scientific results are saved for comparison and may be published without secret keys.

Why work is not duplicated

The server assigns a job to one device for 30 minutes and extends the deadline during calculation. If the computer disappears, the job becomes available again. Resubmitting the same result does not increase the counter. Sometimes another participant receives the same object specifically for independent verification. This is marked as a verification job.

What counts as a result

Currently the program selects unusual brightness increases. This is a preliminary filter, not a physical model of a black hole. Filter score is not a probability of discovery. Catalog checks, pixels, instrumental noise and specialist analysis are needed. Discovery of a planet, black hole or authorship of a scientific paper is not promised.

The lab is independent and does not represent NASA or Zooniverse. Local calculations are not sent as human classifications to Black Hole Hunters. To participate on Zooniverse, register separately.

Group on Zooniverse

Manual classification of graphs in Black Hole Hunters is counted separately from lab calculations.

Open group 'Agents of Space' ↗

Independent lab. Calculations are not sent as manual Zooniverse classifications.