From Feason
Feasy: A Source-Grounded Christian AI Study Companion
Meet Feasy, a Christian AI study companion that answers questions with citations, visible caveats, and tradition-aware theological context.
A confident answer is not necessarily a trustworthy answer. This is especially important in theology, where a short question may carry centuries of interpretation, unfamiliar vocabulary, and genuine disagreement among Christians.
Feasy is our attempt to build an AI study companion that helps without pretending to stand above that complexity.
A companion, not an oracle
Feasy can explain a passage, trace an idea, compare traditions, or help someone find language for a question. It is not designed to issue final judgments on behalf of a church, pastor, scholar, or conscience.
That distinction affects the interface. Feasy reasons in view, attaches citations to claims, and makes room for uncertainty when the available evidence does not support a clean answer. When Christians disagree, the useful response is often a map of the disagreement rather than a synthetic consensus.
Citations are part of the answer
A citation should do more than decorate a paragraph. It should let a reader inspect the passage that is carrying the claim. Feasy's answers connect back to Scripture and the historical sources used to form the response so the reader can examine them directly.
This changes the role of AI. The model is not the destination; it is an interface to evidence. A good answer leaves the reader better equipped to continue without it.
Tradition deserves names
Terms such as grace, sacrament, justification, authority, and tradition do not float free of communities and history. Feasy tries to preserve those contexts. A Catholic source remains Catholic. A Reformed confession remains Reformed. A patristic text is not quietly rewritten as a modern denominational position.
Naming sources and traditions does not remove interpretation. It makes interpretation more honest.
Charity is a product requirement
Theological software can easily reward conflict because conflict produces attention. Feasy is built around a different standard: represent a view strongly enough that someone who holds it would recognize it, distinguish disagreement from caricature, and avoid manufacturing certainty where the sources do not provide it.
This is still difficult work, and we expect the system to improve. But the direction is fixed: every answer should invite examination rather than demand trust.