These answers describe the current compact-first behavior of the extension and the simplest steps to take when a thread still feels heavy.
Compact mode keeps only the newest visible slice of a conversation on screen first. Older messages are hidden behind one inline barrier. This lets the extension reduce pressure from old content while keeping the current part of the chat available sooner.
Load previous messages reveals older history in chunks. The chunk size comes from the popup setting, so you can pull back more context gradually instead of expanding the entire chat at once.
Re-compact returns the thread to the current compact settings after you have revealed more history or temporarily expanded the chat. It is the quickest way to get back to the compact-first baseline.
Show all carefully is the safer full-reveal path. On heavier chats it may use a guard step or restore content progressively. Depending on your settings, the chat can auto re-compact after a short timeout or stay expanded until you re-compact manually.
Start with Compact mode, keep the newest visible count modest, and reload the tab if needed. If the thread contains a large amount of media, files, or embedded content, some remaining slowness may still come from ChatGPT or the browser rather than the extension itself.
Save the settings again from the popup, then refresh the ChatGPT tab. If the tab was already open before a change, a refresh gives the content script a clean pass with the current settings. Also confirm you are on chatgpt.com or chat.openai.com.
Send a short report that includes what you expected, what actually happened, whether the chat was text-heavy or media-heavy, and whether a refresh changed the result. If you use a GitHub issue tracker, you can also open an issue there.