Amos Standard is a study tool. It searches a curated library of Christian texts — Scripture, the confessions, the commentators, the prayer books, the classics — and answers questions from what it finds. It is designed to ground claims in retrieved passages and shows the sources the model cited.
Why it exists
Ask a general AI assistant a theological question and you will get a fluent answer from nowhere in particular. It cannot show you where the answer came from, because it did not come from anywhere: it is the average of everything the model has read, delivered with the same confidence whether it is right or invented. For most subjects that is merely unreliable. For this one it is worse, because the person asking often has no way to check and every reason to want to believe.
There is a second problem underneath it. That averaged answer belongs to no actual church. A Baptist asking about baptism, a Lutheran asking about the Supper, and a Methodist asking about assurance are each handed a blend of positions that none of their traditions hold — smoothed into something inoffensive and, for the purpose of study, useless.
This app is built against both. Its citations open onto the retrieved passages, and the answer is given from within the tradition you told it you belong to. What it will not do is settle for you the questions your church has already settled — the Statement of Faith sets out where it stands and, at equal length, where it deliberately does not.
How an answer is put together
- Your question is used to search the library — by meaning, not by keyword.
- The passages that come back provide the evidence for the answer. The model is instructed to answer from those texts rather than rely on unsupported knowledge.
- The answer is written with numbered markers, and each marker opens the passage it refers to. Sources the answer did not actually use are dropped rather than listed to look thorough.
- When the library has nothing that bears on the question, the model is instructed to say so. That is a valid answer here, not a failure to produce one.
What this buys you, precisely. A citation makes an answer checkable, which is not the same as correct. A model can still misread a passage, or claim more than the text it cites will bear. The remedy is the one the design keeps pushing you toward: when an answer matters, open the source. The source is the authority — the answer is a way of finding it.
The library
The library consists of works in the public domain and material we are permitted to use, prepared source by source rather than scraped: parsed, split for retrieval, and tagged with the author, the work, and the year where we have one. Missing metadata is left blank rather than filled with a guess. Invalid citation markers are removed, but you should still verify that the opened source supports the claim beside it.
What is in it, counted in the database rather than estimated, is on the home page. Its limits are on the Statement of Faith: the source base leans old and leans English, and that shapes what any answer can draw on.
What it is not
It is not a pastor, a counsellor, or a church, and it is built not to behave like one — no personality to perform, no friendship to simulate, no streaks or badges to bring you back. It gives no medical, legal, or financial advice, and it is not emergency help; the Terms of Use say plainly what to do instead if you are in danger.
Who makes it
Amos Standard is made and run by Viktar Danilchyk, one developer based in Poland. There is no company behind it, no investor, and no team — which is worth knowing in both directions: nobody is being sold your data, and nobody is on call at three in the morning either.
It is in beta and free to use: sign in from the home page. Anything else — a correction, a text that ought to be in the library, an answer that went wrong — reaches a person at hello@amosstandard.com.