I wonder how they will _measure_ the text. You can use the range API to get the client rects for individual letters of text nodes in DOM, and then you could ren

I wonder how they will _measure_ the text. You can use the range API to get the client rects for individual letters of text nodes in DOM, and then you could render that into a canvas letter-by-letter. You'd get kerning etc handled by the browser. But canvas has no such API afaik. Maybe Google has implemented their own kerning?