This page describes how Verse Made Simple processes historical texts, generates simplifications, and maintains translation integrity across 31,102 verses.
Source Material
We use the King James Version (KJV), specifically the 1769 standardized text, as our primary foundation. We selected the KJV for three reasons:
- Historical Authority: The most widely printed English translation in history, with unmatched scholarly consensus.
- Literary Resonance: Its language has shaped English for over four centuries and remains a cultural touchstone.
- Public Domain: No licensing constraints, enabling unrestricted computational analysis and redistribution.
We do not process apocryphal texts, modern copyrighted translations (NIV, ESV, NASB), or speculative theological commentary.
The Translation Pipeline
Every simplification is generated through a deterministic, multi-stage pipeline:
- Source Extraction: The canonical KJV text for a given verse is retrieved from our database.
- Structural Analysis: Archaic syntax is parsed and mapped against modern English vocabularies using proprietary prompt engineering templates.
- LLM Processing: State-of-the-art Large Language Models generate the simplified output under strict guardrails that prohibit hallucination, editorialization, or sectarian bias.
- Bottom Line Extraction: A secondary pass compresses the conceptual payload into a single, high-density axiom — the core takeaway stripped of all metaphorical framing.
- Rendering: Final output is wrapped in semantic HTML and delivered via Server-Side Rendering (SSR) for maximum performance and SEO visibility.
The goal at every stage is syntactic clarity, not spiritual reinvention.
Model Selection
We did not default to a single provider or accept whatever the most popular model produced. Before finalizing our pipeline, we conducted a comprehensive evaluation of every major commercially available large language model — testing each against a standardized corpus of theologically complex passages spanning poetry, prophecy, parable, and legal statute.
Each model was assessed in blind comparison across three criteria:
- Semantic Fidelity: How accurately the simplification preserved the original meaning without adding, omitting, or distorting the source text.
- Structural Clarity: How naturally the output read in modern English without sacrificing the gravity of the passage.
- Guardrail Compliance: How reliably the model respected our prompt constraints — resisting hallucination, editorial commentary, and denominational drift.
The model that scored highest across all three dimensions was selected as our production engine. It was not the cheapest option. It was the most precise — and precision is the only metric that matters when processing text that billions of people consider sacred.
Theological Neutrality
Our pipeline is aggressively constrained against bias. We do not inject modern political, cultural, or denominational dogma. The system operates as an objective structural linguist — not a pastor, theologian, or apologist.
However, due to the complex chain of translations underlying the KJV (from ancient Hebrew, Aramaic, and Greek through early English), certain poetic intent and structural nuances may be abstracted during simplification. This is an inherent limitation of any translation process.
Quality Controls
To maintain translation integrity:
- The original KJV text is always displayed alongside the simplification for immediate verification.
- All AI outputs are generated with full context isolation — each verse is processed independently to prevent cross-contamination between passages.
- User-reported errors are reviewed, audited against the prompt constraints, and patched in the rendering pipeline when structural degradation is confirmed.
Report a Translation Issue
If you identify a simplification that materially distorts the meaning of the original text, contact us at contact@kymatalabs.com with the verse reference and a description of the concern.