Research Library / How We Translate It
How We Translate It
Our goal is high-fidelity, study-ready translations you can trust. We combine original-language witnesses, advanced AI research and translation, and manual review.
Original-Language First
For each document, we start from original-language witnesses in our dataset (Greek, Latin, and related traditions where available). We segment and normalize that source text so each section can be studied and translated with clear traceability.
Adaptive AI Research + Translation Pipeline
Each section is translated through a research-aware AI pipeline designed for fidelity and clarity. The research step is intentionally adaptive and may vary by document.
Depending on the text, the pipeline may cross-reference established translations, historical context, technical terminology, and document-specific features before producing the final rendering.
Advanced Models with Manual Double-Check
We use the most advanced AI currently available in our workflow. Every translation is then manually double-checked and compared against other high-quality translations by human reviewers before publication.
Built for Serious Study
This process is designed to maximize accuracy, transparency, and readability so the library can be used confidently for personal study, teaching, and research.
Back to early church documents