CoolFace
Datasetpublic

wrootpress/sermons

Contemporary Preaching: Manuscript Sermons 324 sermons (~665,000 words) written and preached by Wilson Pruitt, a United Methodist pastor, from April 2020 to August 2026 at two Texas congregations (Bee Creek UMC, 166; Covenant UMC Austin, 158). One JSON record per sermon. Why this dataset exists Most public "sermon" corpora are automatic speech transcripts of YouTube video: spoken-word artifacts with recognition errors and no structure. This is the opposite: the… See the full description on the dataset page: https://huggingface.co/datasets/wrootpress/sermons.

sourceHugging Facecc-by-nc-4.0updated 5d agoView on Hugging Face
0likes37downloads
Dataset Card

Contemporary Preaching: Manuscript Sermons

324 sermons (~665,000 words) written and preached by Wilson Pruitt, a United Methodist pastor, from April 2020 to August 2026 at two Texas congregations (Bee Creek UMC, 166; Covenant UMC Austin, 158). One JSON record per sermon.

Why this dataset exists

Most public "sermon" corpora are automatic speech transcripts of YouTube video: spoken-word artifacts with recognition errors and no structure. This is the opposite: the author's own written manuscripts, as prepared for the pulpit, with scripture references, series, and themes. It is meant for work on contemporary preaching, not historical theology.

Fields

id, url (canonical page), title, author, date, church, series, themes (list), scripture (the passages preached), description (short summary), text (full manuscript, Markdown), provenance, licence fields, generated, version.

Provenance and machine involvement

  • —text is the author's own writing. It is not a transcription and has not been drafted or edited by a machine. Manuscripts are copied verbatim from the author's notes; formatting is Markdown.
  • —description and themes were produced with LLM assistance (tagging and summarising); treat them as metadata, not as the author's words.
  • —Only sermons that were actually preached (dated on or before export) and have a real written body are included. Bare lectionary stubs and notes that are only quoted material are excluded by a structural filter. Pastoral, board and funeral material is never in the source repository.
  • —Manuscripts may contain quoted scripture (mostly NRSVUE), quoted authors and illustrations; ordinary quotation only.

Licence

CC BY-NC 4.0, attribution "Wilson Pruitt, sermons.wilsonpruitt.com". Machine-learning training, fine-tuning, evaluation and retrieval, including commercial model development, are explicitly permitted under those terms: keep the attribution and say if you changed the text. Commercial reuse beyond that (print, resale, paywalled products) needs permission: wilson@wrootlabs.com. Full terms: https://sermons.wilsonpruitt.com/rights

Canonical home

https://sermons.wilsonpruitt.com. This dataset is a machine-generated export of that site's build, regenerated from the source and never hand-edited. It is hosted under Wroot Press as a home for texts; the author is Wilson Pruitt.

Loading

python
from datasets import load_dataset
ds = load_dataset("wrootpress/sermons", data_files="data/sermons-2026-09-18.jsonl.gz")