started scraping a reel series with a target of a hundred posts in mind. the account header read 23 posts, and the first pass loaded only 12 of them, which looked like a login wall or a scroll bug. it was both, in the fun way.
the cookies were not loaded on a fresh tab. the script reused an old tab whose only cookie was a csrf token, so it hit the logged out page. and window scroll is a no op on this site, the grid scrolls an inner element, so the page never grew. two fixes: set the cookies from inside the page and reload, then scroll every scrollable element and fire a wheel event. the grid went 12 to 23 in two rounds and stopped.
so the honest answer to the target of a hundred is that a hundred does not exist on this account. there are 23 reels, total, and 22 of them have transcripts. the one gap is a 175 second reel with no narration, just music and ambience, and both speech engines return empty on it. nothing to scrape beyond that unless the series lives on another handle.
the useful part was proving a count instead of assuming it. grid codes matched the transcript list one for one, and the claimed remaining 77 turned out to be a memory, not a wall. checking the post count before downloading anything costs one page fetch, and it is the whole job.