Files are named like Sunday 4th October 2026 - 01h27m43s to 02h15m10s - transcript.txt. The browser saves them to your Downloads folder.
KJV is built in. Matching always uses KJV wording (what preachers quote); the verse is SHOWN in the version you choose.
Any complete 66-book file: .json (3 common layouts) or .tsv/.txt (Book<TAB>Chapter<TAB>Verse<TAB>Text). The file name becomes the menu name. Licensed versions (NIV, ESV…) need your own licence: get the text with tools/fetch_bible.py, then add the file here.
tools/fetch_bible.py
Paste the sermon notes. Every scripture reference is found, in order. Space (or Next) puts the next one on the stage.
Say any of these. Short phrases only work in short sentences so normal preaching never triggers them. A scripture reference in the same sentence wins over navigation. Close variants and run-together words ("netverse") are understood.
Keys (when not typing): Space approve first card / next in plan · X dismiss · 1-5 pick option · ← → or P/N verse · ↑ ↓ chapter · B blackout · C clear · R read aloud · S stop reading · F stage window · / search · H help.
Voice (say it naturally): "John chapter 3 verse 16", "next verse", "previous chapter", "zoom 300", "max zoom", "half max zoom", "blackout", "clear the screen", "approve it", "option two", "read it", "stop reading", "open stage", "close stage", "auto approve on".
Browser limits (honest): speech recognition uses the browser's own service (Chrome/Edge send audio to the vendor's servers; Safari 14.1+ works, Firefox has none). Browsers only allow full screen and opening a window after a click, so "full screen" and "open stage" by voice may be refused: click once in the stage window and press F. "Exit full screen" by voice works. For a fully offline, more accurate setup use the desktop app.