The problem
Glue, foam and paint in a display case can give off fumes that corrode metal objects. Labs test materials, and a "pass" sounds final. It only means the one sample they tested did not hurt the metal in their setup.
For museum display cases
Can this material sit next to that museum object?
A lab "pass" only covers the sample that was tested. Pick your material, object and display, and see what the published evidence really supports, with sources.
Scenario builder
Check a case
Try a case
Evidence disposition
—
Why
What a flat lookup does
Oddy ratings behind this result
Source trail
Knowledge Base context
Agent trace
What this does
It looks up published test results for your material, shows which labs and batches they come from, and says where they disagree or do not apply. Every line links to its source.
Where Sanity fits
Each test, product batch, lab and source is a linked record in Sanity, fetched live. The links are what let it warn that a pass missed a later problem. The answer itself comes from fixed rules, not from Sanity or an AI.
What it will not do
It never says "safe" and never gives a percentage. It only knows what a few published papers report. It helps you think and does not replace a conservator.
Reading a result
Six dispositions. No percentages.
Physical context
Enclosure concentration estimator
Simplified CCI relation C = E × A / (V × N). It estimates concentration only and sets no safety threshold.
Scope
What a result does not mean
“No conflict found” means the scoped evidence did not reveal one. It is not a safety guarantee and does not replace testing or a conservator’s judgment.