Overview
Wmb Table is a desktop study workspace for pastors and Bible students. It brings sermons, Bible reading, notes, subjects, highlights, audio, personal documents, and AI-assisted research into one local-first app.
I started the project in 2020 to help my dad prepare and preach. The first version focused on organizing sermon material. The current version is a much larger tool used by a few pastors for study, sermon preparation, and long-form reading.
Problem
Pastors often prepare messages across many disconnected sources: sermon archives, Bible passages, personal notes, PDF documents, audio recordings, and handwritten topic lists. That creates a lot of repeated searching and manual copying.
The original goal was simple: help my dad find and organize the material he needed before preaching. As the app grew, the harder problem became keeping that workflow fast while adding richer study tools.
Solution
Wmb Table now acts as a structured library and preparation desk. Users open sermons, Bible chapters, and personal documents in tabs, keep reading positions, search sermon content, save passages into notes or subjects, highlight important sections, and sync personal work through Google Drive.
The biggest recent change is the AI assistant. Instead of treating AI as a separate chat box, the assistant uses the content already inside Wmb Table. It searches sermons, reads sermon content, finds Bible passages, works with notes and subjects, and returns clickable references back into the app.
Screenshots
Reading view with sermon tabs, translation controls, library navigation, and the right-side content browser.
Search results with highlighted matches and grouped sermon references.
Subjects organize passages around themes for sermon preparation and study.
Notes keep study material separate from the reading surface while staying close to the library.
Highlights work across Bible passages, sermons, and personal documents.
AI chat answers with Bible and sermon references linked back into the app.
Audio playback and text tracking help users listen while studying.
Technical Approach
The app is built with Electron, React, TypeScript, and Ant Design. The main process owns the database, file system, indexing, document import, sync, AI tools, and audio services. The renderer focuses on reading, search, navigation, and editing workflows.
The data layer moved from the older local document-store approach to SQLite. Reference data is bundled read-only, including sermon metadata, language packs, translator data, and audio links. User data lives in a separate SQLite database managed with Drizzle migrations. Bible versions are stored as separate SQLite files with FTS5 search.
Search uses Tantivy for fast sermon lookup. The app also keeps local Bible search through SQLite FTS5. Google Drive sync tracks changes for notes, subjects, suggestions, and custom documents, so users keep their personal work across machines.
The AI layer uses the AI SDK with provider support for OpenAI, Anthropic, Google, and OpenAI-compatible models. The assistant has domain tools for sermons, Bible passages, notes, subjects, languages, and personal documents. This keeps generated answers tied to the user's study library instead of generic model output.
Key Features
- Sermon library with 1000+ sermons and multilingual translations
- Full-text sermon search with match previews and deep links
- Bible reader with multiple local and downloadable versions
- Persistent document tabs and reading positions
- Notes with sermon and Bible references
- Subjects for grouping passages around a theme
- Highlights across sermons, Bible chapters, and personal documents
- Custom PDF document import
- Google Drive backup and sync for personal work
- Audio playback and text-to-speech support
- AI chat with app-specific tools and clickable references
- Auto-update support for desktop releases
Challenges
The main challenge has been balancing depth with speed. Sermon study involves long documents, multiple languages, and many cross-references. The app needs to open quickly, search large local datasets, and keep user-created data separate from bundled reference data.
AI added another constraint. The assistant must help without losing the source trail. I handled this by giving the assistant structured tools and reference links, so answers point back to sermons, Bible passages, notes, subjects, and documents inside the app.
Sync also required a careful data boundary. Reference material stays read-only. User work is tracked separately, so backup and restore focus on the data that matters most to the person preparing a message.
Outcome
Wmb Table has grown from a personal app for my dad into a published desktop tool used by a few pastors. The project is still active, and recent work has focused on making AI useful inside the actual study workflow rather than placing it beside the workflow.
The current version is no longer only a sermon organizer. It is a local-first preparation environment for reading, searching, listening, note taking, topic building, backup, and AI-assisted research.
Tech Stack
- Electron 40 and Node.js 20+
- React 19 and TypeScript
- Ant Design, Zustand, Framer Motion
- SQLite, better-sqlite3, Drizzle ORM
- Tantivy for full-text sermon search
- SQLite FTS5 for Bible search
- Google Drive API for backup and sync
- AI SDK with OpenAI, Anthropic, Google, and OpenAI-compatible providers
- Sherpa-ONNX, Edge TTS, OpenAI TTS, ElevenLabs, and Cartesia integrations