Most pastors and preachers carry the same problem. The sermon needs to be ready by Sunday, and the prep involves exegesis, commentary reading, original-language work, illustration hunting, drafting, rehearsal, and delivery — all on top of the rest of the week's pastoral responsibilities. The tools for each part are different. The notes are scattered. By the third year of preaching weekly, the same passages come around again, and the prep has to start over because last time's work isn't findable.

This guide is about putting the parts of preaching prep that benefit from being together — exegesis, illustrations, past sermons, study notes — into one vault where an agent can read across them. The actual study still happens. The hunt for what you've already done doesn't take half the week.

## What sermon prep actually has to hold

Preaching prep over a long ministry develops repeating elements. The same passages come around (lectionary churches every three years; topical preachers more variably). The same theological themes recur. The illustrations you've used start to feel reused. The exegetical work you did three years ago on a passage is somewhere in a folder you can't find.

Four loads a preacher's vault has to carry:

- **An exegesis file per text** — the original-language work, the commentary notes, the historical context, the questions the passage raises.
- **A sermon archive** — every sermon you've preached, with the manuscript or outline, the date, the text, and the post-sermon notes.
- **An illustration library** — stories, examples, quotes, and analogies you've collected. Tagged by theme so the agent can pull the right one for the right text.
- **A teaching and study log** — what you've read, what you've taught, what you're still working through.

A vault holds all four together. Plain markdown for the text. PDFs (commentaries, theological articles, monographs) auto-converted to searchable text. Audio (sermon recordings, study group meetings) transcribed with speaker labels. The same shape works for any teaching practice that builds up over years; see [AI notes for course creators](/guides/creatives-content/course-creators-curriculum-design/) for a structurally similar setup with a different application.

## The exegesis file — one page per passage, deepening over years

The core working asset is a page per passage you've preached or studied. Not the sermon — the underlying exegetical work. The page accumulates over years. The first time you preach on a text, the page is shallow. The third or fourth time, it's a real reference.

A typical passage page holds: the text in your preferred translation (or several), original-language notes, structural observations, key word studies, the questions the text raises, the interpretive options you've considered, the commentaries you've consulted, and links to past sermons you've preached on the passage.

When the lectionary brings the passage back around three years later, you open the page first. The agent helps with the comparison: "I'm preaching on Mark 4:35-41 again this year. Pull what I noted last time. What questions did I raise that I didn't fully answer? What angle did I preach last time that I might want to take a different approach to this year?"

You start the prep with three years of accumulated thinking, not from scratch. The blank-page problem evaporates because the page isn't blank.

For commentaries — most are dense PDFs nobody re-reads — drop them onto sub-pages under the passage. They auto-convert to markdown via docstrange so they become searchable text. When you're stuck on a particular phrase, the agent can pull what each commentary says about it: "From the commentaries on this page, find what each says about the disciples' question 'do you not care that we are perishing?' Summarize the interpretive options."

## The original-language work and the agent that reads it

Original-language exegesis is where most working pastors save the most prep time. Not because the agent does the exegesis — it doesn't, and it shouldn't. But because the agent handles the surrounding work.

Type the original text into the page. Add your translation notes, the morphology observations, the syntactical questions. Drop in PDFs of grammars, lexicons, and word-study sources. PDFs auto-convert to markdown so the agent can search across the linguistic resources you've gathered.

When you're working on a difficult phrase, the agent helps with retrieval: "From the lexica and word-study resources on this page, pull what each says about the term used for 'rebuke' in this passage. Note the semantic range and the parallel uses elsewhere in the New Testament." You get a starting point. You verify against the actual sources. The judgment is still yours.

For Greek and Hebrew, the agent's value is in the cross-referencing and the surfacing. The translation, the parsing, the interpretive call — those stay with you. Use the vault as a working tool, not as an authority on what the text says.

## The sermon archive — every manuscript, dated and findable

Every sermon you've preached gets a page in the archive. The manuscript or detailed outline. The date. The text. The audience or congregation. The illustrations you used. The post-sermon note about what landed and what didn't.

The archive is the asset that compounds. After five years of weekly preaching, you have 250+ sermons. After fifteen years, you have 750. Most of them are searchable text the agent can read across.

When you're working on a new sermon and you remember vaguely that you preached on a related theme before, the agent finds it: "Find every sermon I've preached in the past five years that addresses the theme of suffering and trust. Pull the texts and the central angles." You get a real synthesis of what you've already preached, which prevents accidental self-repetition and reveals the patterns in your own development as a preacher.

For seasonal or lectionary preachers, the archive becomes a year-by-year record of how you've handled the same texts. Useful for the obvious comparison, also useful for spotting where your preaching has shifted in ways you didn't notice in real time.

## The illustration library — stories that compound, not stories that get reused

Most preachers have a small set of illustrations they reach for. The personal story. The historical example. The literary reference. The current-events tie-in. Most carry these in their head and lose track of which they've used in front of which congregation.

An "Illustrations" sub-vault, with one page per illustration, captures them. Each page has the illustration written out, the lesson it makes vivid, tags for the kinds of texts it suits, and a log of the sermons where you've used it.

When you're drafting a new sermon, the agent reads the library: "Pull every illustration I've used to make the theme of grace concrete. For each, note when I last used it in this congregation and how it landed." You get a working list with the institutional memory that prevents you from telling the same anecdote to the same congregation a second time.

For illustrations you've decided not to use anymore — the one that didn't land, the one that aged poorly, the one that you've come to question theologically — archive with a note explaining why. The next time you draft on a related theme, the agent reads the archive note and skips the illustration. You don't accidentally bring back the story that's been retired for good reason.

## The drafting flow — outline, draft, rehearsal

The drafting flow varies by preacher. Manuscript preachers work from full text. Outline preachers work from headlines. Extemporaneous preachers work from a few cards and trust the Spirit. The vault supports any of them.

For manuscript preachers, the draft lives on the sermon page as a markdown document. The agent helps with the structural work: "Read the draft. Where does the sermon slow down? Are there sections that feel told rather than shown? What's the strongest moment, and is it landing in the right place — beginning, middle, or end?" You get observations to react to.

For outline preachers, the page holds the outline with sub-bullets for each movement's intent. The agent helps with structural pacing: "Each movement is roughly how long? Where am I underweight on application versus exposition? Which transitions feel forced?"

For the assembly part of drafting — pulling the relevant illustration from the library, the cross-references from the exegesis, the historical context from the commentary work — the agent does the boring work. You do the writing that makes the sermon yours and makes it suit your congregation.

## Rehearsal and the post-sermon debrief

Reading the sermon out loud — even just once — catches the parts that work on paper but not when spoken. The pattern: rehearse out loud, record on your phone, drop the audio onto the sermon page. It transcribes with speaker labels.

Now the rehearsal is part of the page. You can ask the agent: "In yesterday's rehearsal, find the moments where I stumbled. Pull the sections where the wording felt awkward." You get the relevant moments back, in plain text. You revise. The Sunday morning delivery is calmer because the awkward bits got caught the day before.

After the actual sermon, type three lines on the page. What landed. What didn't. What you'd change. Two minutes of typing. Over the years, those debriefs become some of the most useful pages in the vault. They're the only place where you've honestly recorded what worked in front of an actual congregation versus what you assumed would.

The same capture-after habit is the spine of [how to capture action items so they actually get done](/guides/meetings-people/action-items-actually-get-done/) — different output, same shape.

## Bible study and small group teaching

The same vault supports the rest of the teaching ministry. Sunday school lessons, small group studies, mid-week teaching — all of it gets a page. The exegesis from sermon prep feeds the small group study. The illustrations from the sermon library feed the teaching. The questions the small group raised feed the next sermon's prep.

For small group meetings specifically, drop the recording onto the meeting page (with participants' knowledge and consent). Speaker labels mean you can later remember who raised which question. The agent can read across the small group records: "What questions has this small group raised about discipleship over the past year? Have any recurring themes emerged that might be worth a series?"

For the teaching that involves outside materials — books, articles, sermons by other preachers — drop the PDFs in. They become searchable. You stop forgetting which book had the chapter on the topic you're teaching.

## A boundary and a humility note

A practical note: the agent is doing assembly, retrieval, and structural drafting. It's not doing exegesis. It's not doing theological reasoning. It's not pastoring. The sermon's content, the interpretive judgments, the application to the congregation — those stay with the preacher.

The vault is a working tool. It makes the prep faster and the institutional memory deeper. It doesn't replace prayer, study, the life of the preacher, or the work of the Spirit in preaching. Use it for the chores. Do the actual work yourself.

For pastors managing the broader pastoral and administrative side alongside preaching, the parent guide [AI notes for church leaders](/guides/creatives-content/church-leaders-sermons-pastoral/) covers the wider workflow.

## A calmer way to prep

Sermon prep doesn't have to feel like a Saturday-night cram. The cram comes from doing all the assembly work the day before. The shape that holds is small ongoing capture (drop new commentaries, illustrations, and study notes into the vault as you find them) plus a focused weekly block where the agent's first-pass synthesis is already done and you're doing the second pass that requires your judgment.

Try Docapybara free — [sign up](/accounts/signup/), pick the text for next Sunday, drop in your translation notes and one commentary section, and ask the agent for a list of past sermons you've preached on related themes plus a draft outline grounded in your existing material.