Reach, not enclosure
A study nobody can get to has failed regardless of how good it is. The public-domain layer is free forever, a real portion of our own writing is free forever, and search and citation lookup stay open across the whole corpus.
Carrying the work to the people who want it.
Divine Data is the product of Sophia Partners AI. It is one body of work served three ways: a library you can read, a query engine you can ask, and an ebook you can simply buy. Behind all three is a single commitment — never fabricate. Every statement is grounded in a source, carries the reference it came from, and is labeled for what kind of claim it is. This page explains that commitment in plain terms, says how we price, and is honest about how the work gets made.
The mission
Outreach and the transfer of knowledge
Put plainly: shine light on the light. Give people the ability to see scripture's architecture and symbolism, its connections to the literatures history left out, and — eventually — the ability to query across all of it at once, synthesized and cited rather than taken on faith.
That's the whole reason the project exists, and it is written down as binding internal policy, not as marketing. It means we carry this body of work to the people who want it and try to make it genuinely understandable — rather than hoarding it, gating it behind gimmicks, or diluting it into something easier to sell.
Understandable, not simplified away
Verse-by-verse commentary in modern English, plain-language questions answered from the corpus, and a plain-English rendering alongside the translation. Clarity is the point; flattening the difficulty is not.
Trustworthy, or it is worth nothing
Transferring knowledge only counts if what arrives is true and checkable. That is why the standard below governs the writing itself, not just the software.
Divine Data sells trustworthy access to a unique body of knowledge, not tokens.
The Prime Directive
Never fabricate — and always show your work
This is our first commitment to you, and it is the one we would rather ship late than weaken. It applies to everything with our name on it: the verse-by-verse studies, the essays, the translation notes, the ebook, and every answer the query engine gives. There is no separate, looser standard for the writing and a stricter one for the software.
In practice it means every user-facing statement is exactly one of three kinds — and it tells you which one it is.
What the source text says on its face. Quoted, not paraphrased into something it does not say.
A claim tied to a specific place in the corpus — a verse reference, a commentary location, a named document — so you can go and check it.
This project's Gnostic and Pistis Sophia reading, marked explicitly as ours. Never presented as fact and never laundered into scholarly consensus.
There is no fourth kind
The thing we refuse to produce is an unanchored assertion dressed up as fact or as our research. Four habits follow directly from that refusal:
- Where the sources differ, we show the variants rather than silently choosing the reading we prefer.
- We explain differences plainly — without polemic, and without pretending a disputed point is settled.
- We correct the record visibly the moment we find we were wrong, instead of quietly editing and moving on.
- We say when we do not know. If a question falls outside what the corpus covers, the honest answer is that it is outside the corpus — not a plausible-sounding guess.
Why this matters for an AI-assisted product specifically: a language model can produce a fluent, confident paragraph about any of these texts from its training data. That paragraph is not our work, cannot be traced, and cannot be corrected. An answer built that way and presented as corpus content is a violation of this standard, full stop — which is why answers are assembled from the corpus and carry their references with them.
How we price
Fair, legible, and never designed to trap you
We intend to make money — the project has to be sustainable to keep serving people. But profit is held to the same written standard as everything else: fair, stated up front, and never extracted through confusion.
Prices, what is free, and how settled each figure is are all laid out there in detail — including the parts that are not final yet, marked as not final.
- No dark patterns. No manufactured urgency, no countdown timers, no confusing toggles, and no roach-motel cancellation. Leaving is as easy as joining.
- A genuinely useful free path. Not a crippled demo, not a trial that expires. It is real and it does not shrink after you sign up.
- No nickel-and-diming. Honest caps by plan, stated in advance, failing closed at the cap rather than surprising you with a bill.
- Public-domain material stays free. The translation and the modernized scripture text are not ours to sell, so we do not sell them.
- The price is published before the checkout exists — except the ebook, where checkout already exists and works today.
How this is made
A human-led, AI-assisted project — stated plainly
Sophia Partners AI is an AI company, and we would rather tell you what that actually means than let the word "team" do misleading work. Divine Data is built by a human-led, AI-assisted process. Chris is the founder and CEO; he sets direction and holds every build approval. The rest of what we internally call "roles" — product, design, engineering, editorial, and the review functions — are AI agent capability descriptions, not employees. They are ways of organizing the work, not people.
No invented staff
You will not find a "meet the team" wall of faces and bios here, because it would not be true. One human founder, plus AI capabilities working under his direction, is the accurate description — so it is the one we publish.
Nothing ships unapproved
Ideas are proposed continuously; building waits for the founder's explicit go-ahead. That approval rule is written down internally and it is what keeps an automated process from quietly publishing in our name.
Two vetoes that outrank speed
Two internal review functions can stop work outright: one enforces the truth standard above, one enforces rights, privacy, and cost limits. Neither can be overridden to ship faster or to make a dollar. When they conflict with a deadline, the deadline loses.
What we are deliberately not telling you
Because there is nothing true to say yet. A page like this is usually where a company reaches for social proof; we would rather leave the space empty than fill it with invention. So, explicitly:
- No user or customer numbers. The hosted query product is not open yet, so any figure would be fiction.
- No funding or investor claims. We are not making any.
- No press mentions, awards, or testimonials. If we ever have them, they will be real and linkable.
- No founding story or company timeline. Beyond the fact that Chris founded and leads it, we are not dramatizing a history.
- No headcount. See above — the honest number of human employees is not something we will inflate with agent names.
If any of this changes, this page changes with it, and the claim will arrive with something you can check — the same rule the rest of the corpus lives under.
The next step
Three doors in — read it, ask it, or buy it
The mission is only real if the work actually reaches you, so the best thing this page can do is hand you off. The library is complete and readable today. The query engine is built and eval-tested, and runs locally while the hosted app is being finished. The ebook is a real, working purchase right now.
Read the corpus
All 39 Old Testament books studied verse-by-verse, the full 148-chapter Pistis Sophia translation with a plain-English rendering, and the essays built on top of them.
Ask the corpus
Divine Data Answers takes a plain-language question and builds the reply only from the corpus — cited by reference, tier-labeled claim by claim.
Buy the ebook
The finished, produced Pistis Sophia edition — two formats, real checkout, instant download.
Contact
Questions, orders, or something wrong
For anything about an order, a download, or this site generally, reach out at hello@example.com.
Placeholder address — replace with a real support inbox before this site goes live on a real domain.