planned
Rate a proposal against a configurable rubric.
An LLM that ingests a proposal, scores it against your rubric, and writes a defensible critique. The point isn't replacing human judgment — it's spending human attention where it matters. With 200 inbound proposals and a quarter to evaluate, triage with grant_scorer in an afternoon, then deep-dive on the top twenty.
Most funders already do something like this manually. Making it cheap shifts the equilibrium: smaller funders get the same triage capacity as the largest ones, and the marginal proposal stops getting filtered out by attention scarcity.