GPTBot downloads your JavaScript. It does not run it.
Over three weeks, OpenAI's training crawler fetched our Next.js bundle page by page, and the React code in it never ran. Its other agents, the search crawler, the robots.txt readers and ChatGPT-User, fetched no JavaScript at all. A request for a .js file is a download, not a render. The same goes for llms.txt: the first verified OpenAI fetch of ours came a day after a page linked to it.