An Architecture for Grant Discovery and Grounded Application Drafting

Grant seekers face a discovery problem: tens of thousands of private foundations give billions of dollars each year, little of it through open calls, yet the usual tools only index the calls. Grantivo is built on a different premise: the most informative signal about a funder is its record of past giving, disclosed in annual public filings.

The paper describes how those filings become a queryable model of funder behavior, more than 140,000 foundations and over 7.5 million grants, and how that model powers discovery, funder research, evidence-backed drafting, and draft audit. It also examines the property we find most interesting: because the writing happens inside the system, it observes the applicant-side text and outcomes that the filings omit, supervision no public dataset contains, together with the data governance that using customer applications responsibly requires.