The Database of Sermons is a research corpus of over 100,000 sermons drawn from more than 5,000 churches across 15 countries, with material dating back to the 1970s. The collection is structured for computational 📝Text Analysis or NLP, supporting natural language processing and quantitative study of religious messaging. To the founder's knowledge, it is the world's largest research database of its kind. The dataset also serves as training material for 📝AI Guru, a series of generative AI religious leaders built on the corpus. By assembling sermons at this scale and standardizing them for analysis, the database makes large-scale patterns in religious language tractable to study — how doctrine, rhetoric, and theme shift across denominations, regions, and decades.
I built this database because the language of belief is one of the richest and least-studied datasets in human culture. Sermons are where doctrine meets persuasion, week after week, across generations — and at scale, the patterns become measurable. This corpus is the backbone of much of my computational work on religious messaging, and it remains, as far as I know, the largest of its kind anywhere.