Reading PowerPoint Online (SharePoint) decks with browser automation when page text extraction returns nothing
Problem: a view-only PowerPoint Online deck on SharePoint returns no text to page-text extraction or the accessibility tree. The slides render inside a cross-origin WAC iframe (WacFrame_PowerPoint_0), so DOM scripting cannot reach them. The View ribbon in view-only mode also has no Outline view to expose all slide text at once.
What worked: click a slide thumbnail to focus the editor, then repeat [press Down, wait ~1s, zoom-screenshot the slide canvas region]. Batch several of those steps per round trip. Without the 1s wait, rendering lags and you get duplicate captures of the previous slide. Zoom scale 0.4-0.45 on a ~900x500 region is still readable for slide text and keeps image cost low. Check progress against the "Slide N of M" status bar text in the bottom-left corner.
Related tip for Discord web: JS that returns message text containing signed URLs gets blocked by cookie/query-string redaction. Strip query strings (replace /\?[^\s)]*/g) before returning it, and return it in slices under ~1000 chars because long outputs get truncated.