Your entire content workflow, streamlined.

One caption system with three styles. Font: Inter ExtraBold (800), tracking about -4% (the closest free match to the heavy, tightly set grotesque here), lowercase as spoken, punctuation kept. Words appear 1 to 3 at a time (usually 1 or 2), each group cut on and off with the words, with no animation, no highlight on the spoken word, no outline and no shadow. 1. Full-frame talking shot: white #FFFFFF, about 3.5% of the frame's height (~68 px), centered, sitting over the chest about 64% down, below the chin. 2. Split layout (graphics canvas above, speaker window below): about 3.1% of the frame's height (~58 to 62 px), centered, about 66% down, in the gap between the graphics and the speaker's head, about 40 to 60 px above the window's top edge. On the light gray canvas it's plain near-black #111111. When it would land on graphic content (a card, a comment row, a pixelated item), it stays readable as white text on a translucent dark pill instead: black at about 40% (reads #848484 over the canvas), about 74 px tall, corner radius ~16 px, ~18 px padding left and right. On the black canvas it's plain white. 3. Emphasis lockups, on full-frame shots only, about one every 20 to 30 s, on the word that carries the line (the named method, the key adjective, the turn): a) a single key word set in Instrument Serif Italic instead of the sans, white, same size or a touch larger; b) a term lockup: a small Instrument Serif Italic word (about 2% of the frame's height, ~40 px) over a big Inter ExtraBold all-caps word (about 5.7%, ~110 px), built word by word as spoken, the big word coming in gray and brightening to white over ~4 frames, held until the next line; c) the closing call to action: the call-to-action verb in the sans with the keyword to type on the line below in Instrument Serif Italic inside curly quotes, about 4% of the frame's height (~80 px), white, centered over the chest. No blend modes: every caption is a solid fill.