Reference programs

Each program answers one narrow, pre-declared question, and each one reports its own state.

Program	Material	State in this release	What that means
Exogenous insulin	rapid-acting analogue (SC), regular human insulin (IV)	Runs	Paired control and treated arms per member, from an identical starting state
LBHR-030 Greater burdock	arctigenin	Margin computable	Modeled exposure against the measured active range. Exposure comes from published piglet pharmacokinetics scaled allometrically, so it is not admissible as human exposure
LBHR-135 Bilberry	cyanidin-3-glucoside	Runs at the studied dose only	Measured human exposure exists only for 500 mg
LBHR-172 Lion's mane	erinacine A	Exposure only	Active range not admitted
LBHR-129 Yerba santa	sterubin	Blocked, reason named	No admitted pharmacokinetic source
LBHR-062 Rosemary	carnosic acid	Blocked, reason named	No admitted pharmacokinetic source

Validation status: 0 of 6 validated in vivo

Running a program is not validating it, and nothing here is dosing guidance. livemore programs list reports "0 of 5" because it counts only the five botanical programs.

A purified compound is not treated as interchangeable with a plant extract, a different plant part or a metabolite mixture.

Running a program

livemore programs list                       # what can and cannot run, and why
livemore programs run insulin --cohort 12 --seed 0 --json
livemore programs run LBHR-135 --dose-mg 500

Options: --dose-mg, --cohort, --seed and --json. The same programs run from the interactive shell (program run LBHR-030 dose_mg=400 cohort=10 seed=1 output=run.json), the Python SDK and the REST API.

The refusals are part of the product

When a program cannot answer from admitted evidence, it says which measurement is missing instead of producing a plausible-looking number. Test this deliberately: ask for a dose nobody measured, or a compound with no pharmacokinetic source. See worked examples for real refusals and the questions the evidence can answer instead.