Gurutva how much of that picture is real?
Download  ↓

We buried one block. Your survey cannot find it.

On the left is what is actually down there. We planted it, so the answer can be graded rather than argued about. The four beside it are what your readings still allow — every one of them reproduces the same measurements, to within twice the wobble of the instrument. Ordinary software picks one of them and draws it for you, in colour, with no warning.

What is really there
Possibility 1
Possibility 2
Possibility 3
Possibility 4
All four sets of readings, on top of each other. They differ by at most 0.111 — about twice the wobble of the instrument — while the pictures above differ by 0.53 g/cc, more than the entire body being hunted. Tiny differences up here, enormous differences down there. That is what makes this hard, and it does not go away by buying a better meter.

None of the four looks like the block. That is not a bug in the software — it is what your data left undecided. Which is itself measurable, and almost nothing in this industry reports it.

Download for Windows  ↓ 71 MB. Double-click it. Runs on your laptop; your survey is never uploaded anywhere.
No Python, no install, no admin rights. The first time you open it, Windows asks you to confirm: click More info, then Run anyway, and it starts. Prefer to read the source first? The 65 KB Python version is the same code, and needs numpy scipy matplotlib.

What buying more survey actually buys you

dense rock light rock faded = the software is guessing here • sensors
4 — cheap30 — thorough
That range is the product. Drag the slider: the picture does not change much, but how much of it is real does. Right now the data actually constrains of it — the rest is the software filling in.
Every number here comes from the real engine, on a scene where we planted the answer, so the checks can be graded honestly.

Three things it is allowed to say

NOT CLAIMED

A check failed. The numbers are diagnostics. Do not put them in a decision.

PROVISIONAL

Every check passed, but nothing was compared against a known right answer — because real data has none.

CLAIMABLE

Every check passed and it was verified against a known answer. Rare, and meant to be.

-346 Mt
Utah FORGE · real DOE data
95% between -484 and -207. The standard workflow would have called this interval 2.6× tighter than it is. Read the whole report →
72%
CO₂ containment
A 400 m leak two kilometres out is only detectable above 72% saturation. Below that, no containment claim covers it — and we say so instead of quoting a number. Read the whole report →
14.6σ
A dark-matter halo
The same code, unchanged, at 1022 times the distance. All four checks passed and the fifth caught it anyway, so the map is not claimed. Read the whole report →

Those are the actual documents the tool writes, not summaries of them. The gate rows, the refusals and the figures are exactly what lands next to your data.

Who this is for

You are about to drill. Get the range, and the depth below which the picture is invented, before the rig moves.

You store CO₂ and have to prove it stayed put. Get a defensible statement of what your monitoring can and cannot show — before you fund it.

You want to check the work. 175 passing tests, and the retractions are in there too. The code is public, under Apache‑2.0.

How you actually use it

  1. Download Gurutva.exe and double-click it. There is no step two: no Python, no install, no admin rights.
  2. Press Use the example, then Run, to watch it work on a survey that ships inside it.
  3. Then choose your own CSV and answer three questions about your site. We do not guess them for you: they are statements about your ground, not ours.
  4. A report opens in your browser, saved next to your data.

Download Gurutva.exe  ↓

There is no sign-up, no upload, and no server. It runs on your laptop and your data never leaves it. If a check fails it will tell you so and refuse to give you a number — that is the product, not a fault.

Tell me what broke — github.com/vaishak-v-nair/Gurutva/issues