Why VisiGrid exists

The problem

Spreadsheets are the most trusted tool in business — and the most dangerous.

Silent failures. Opaque calculation. No audit trail. Not because they're broken. Because they were never designed to be verifiable.

A model runs. The output looks right. Someone changes an input. The output still looks right — but is it? You can't prove it. You can't even know what recalculated.

Entire industries build critical decisions on this foundation. They know it's fragile. They accept the risk because there's no alternative.

Our thesis

Computation should be deterministic, inspectable, and verifiable.

Same inputs. Same formulas. Same outputs. Always.

If a spreadsheet recalculates, you should know it happened. If a value changes, you should know why. If someone hands you a file, you should be able to prove the outputs match the inputs.

This is not a new idea. It's how compilers work. It's how build systems work. It's how every serious engineering tool works.

Spreadsheets just never caught up.

What we optimize for

Determinism over convenience. You control when values update. No silent recalc. No ambient state leaking into your results.

Auditability over features. Every cell knows what feeds it. Every output can be fingerprinted and verified. The computation is the proof.

Trust over polish. We'd rather ship fewer features that work exactly as documented than many features that mostly work.

What we don't do

No hidden recalc. No ambient state. No features that compromise verifiability.

If a function needs randomness, network access, or the current time — it can't run in VisiGrid. That's not a missing feature. That's the design.

Custom functions run in a Lua sandbox with hard instruction limits. No globals. No I/O. No side effects. If it runs, it runs the same way every time.