Slides That Follow You

Present without a clicker. A speech model and a small embedding model on your laptop listen to what you're saying and move to the slide you've started talking about.

Models not loaded. Whisper (~80 MB) and MiniLM (~23 MB) download once, then cache.

Markdown, one slide per section, separated by a line with ---. Use # for the title, ## for a subtitle, - for bullets, > for a quote. Add cue: lines with phrases you're likely to say; they help matching and aren't shown.