My theory is that they rate limit that URL aggressively due to AI scrapers. At this point it's faster to just clone the repo and do your searching locally.