I spoke to the original creator of Google Docs (from before it was Google Docs, https://www.theverge.com/2013/7/3/4484000/sam-schillace-inte...) and he confirmed what you recall; they had to write their own text rendering code because the browser does not implement the behavior required by a word processor (one example: flowing text around an image)
That was long before they switched to canvas rendering, too: they were still just using contenteditable, though potentially with mild layout augmentation with things like absolute positioning and sizing of things (I don’t know).
Flowing text around an image: the web has been able to do this for donkey’s years, it’s the float property. More recently, there’s the CSS shape-outside property that even lets you use non-rectangular crop: https://developer.mozilla.org/docs/Web/CSS/shape-outside.
IIRC float is just left or right, you can't embed an image "in the middle" of text. I'm no web expert and the conversation was a decade ago, so I'm sure something got warped in translation.
Ah, in the middle. Yeah, you still can’t do that. There was https://www.w3.org/TR/css3-exclusions/ which let you do it, but it was only ever implemented in IE 10–Edge 18 (prefixed), and seems dead now.
