Software establishes the maths
A deterministic engine decides every mathematical result. AI proposes interpretations; you confirm meaning. These are separate kinds of evidence, never one “AI verified” score.
About
Rules decide who gets in, who qualifies, and what happens next. Tangibool exists so the people who write those rules can see exactly what they do, before anyone else finds out the hard way.
A rule can be perfectly well formed and still wrong. It can group conditions the wrong way, forget an exception, or turn a necessary condition into a sufficient one. A diagram alone may not show the failing case, and a plausible explanation isn’t evidence.
We think the gap between a written rule, a symbolic expression, and what it actually does across every case should be visible and fixable, without anyone first having to become an expert in formal methods.
So Tangibool puts the words, the formula, and the diagram on one model, checks every case exactly, and tells you plainly what it did and didn’t establish.
“Allow entry to members and invited guests, unless they are banned.”
(member OR invited) AND NOT bannedmember OR (invited AND NOT banned)What we believe
These aren’t marketing lines. They are design rules the product is built and tested against.
A deterministic engine decides every mathematical result. AI proposes interpretations; you confirm meaning. These are separate kinds of evidence, never one “AI verified” score.
Every result names the property, the assumptions, the method, and the revision. “Not determined” is an honest answer; a timeout is never a pass.
Local first, cloud by choice, AI only with consent. No product analytics, and no formulas, notes, prompts, or names in our logs.
Keyboard alternatives to every drag, labels for screen readers, truth values that never rely on colour alone, and four content languages.
Deterministic exports in open formats, full-history packages you can import again, and a headless runner that works without us.
Plans differ in cloud limits, never capabilities. Lessons are open to all, viewers and commenters are free, and nothing is deleted because of a plan.
How it’s built
The same deterministic logic engine runs unchanged in your browser, on our servers, and in the command-line runner, so a result means the same thing wherever you see it.
It pairs exhaustive enumeration with an independent SAT solver, and every witness is replayed in the evaluator before it reaches you. Before a release, randomized fixtures are run through all three and must agree without a single mismatch. Generated Python, PostgreSQL, and SMT-LIB are executed and compared with the engine, too.
Exact results are exact within documented bounds, and when something goes beyond them, the product says so.
Who it’s for
Who want an exact equivalence result, or a reproducible counterexample, before refactoring a conditional.
Who need investment restrictions and pre-trade rules stated precisely, checked exactly, and kept with a reviewable history of every change.
Who turn written eligibility, access, and workflow rules into logic with a confirmed interpretation and independent tests.
Who want worked examples and explanations tied to real evaluation steps, not illustrations.
Company
Questions about the product, a course, your team, or your data? We’d like to hear from you.
Tangibool is operated by Foreranger LTD.
Eastgate House, Dogflud Way, Farnham, Surrey, England, GU9 7UD
Registration: 10535686
Registered in England and Wales
VAT number: 258669742
General enquiries: [email protected]