About
Yablessom is one home for three things that usually live apart.
Being builtBack to the DoorwayWORD FOR WORD
Our AI never types a verse. Code pulls it from the text, word for word. A 2026 study by The Apologist Project, with Fide AI, tested that method and found it exact 100% of the time, across all 10 models and every English translation it tried. The code guards the words. The community guards the meaning. Bible Weapon itself wasn't part of that study and has no answering AI yet, so we'll publish our own benchmark.
Source: The Apologist Project, "Scripture Quotation in Generative AI", 5 August 2026. The study comes from an advocacy group working with an AI lab. The 100% is for English translations: in the other languages it tested, the same method scored 99.4%. "Exact" counts every character except spaces, punctuation included.
WHAT IT TOOK
What it took, so far.
We think you should see what this costs. So here is Bible Weapon's build ledger: the time since the first working app, and the AI used to build it, each figure with the date it was counted. One honest caveat. The token count measures the AI's work, not ours. The hours people put in, reading, testing and deciding, aren't in it.
WHAT IT TOOK: BIBLE WEAPON
- First working app
- 25 July 2026The idea and the planning started earlier. That date isn't on record, so the ledger starts at the first working app.
- Time since then
- 3 months (25 July to 29 October 2026)
- AI tokens used
- 1,407,160,740 (to 20 October 2026)
- API-equivalent value
- about $686 (to 20 October 2026)An API-equivalent estimate, not what we were billed. We build on a subscription.
- Code commits
- 46 (at the audit, 20 October 2026)
The token count doesn't capture the hours. When we have a real count of them, it goes here.
These figures are Bible Weapon's only. The website, the agents and the rest of Yablessom aren't counted yet.
Every figure shows the date it was counted, and changes only when we re-count.