What would you like Inkpour to do? and Why is this useful?
"When a conversation in ChatGPT gets very long, it adds a kind of table of contents on the right next to the scrollbar (attached image), where it summarizes the user's question into a short heading. By hovering over this table of contents, you can quickly navigate through long chats.
Since ChatGPT already generates this, we could add the exact same table of contents at the top of the exported file, allowing users to click on these short questions to jump to them within the exported document. This means we already have all these headings inside the exported file (Markdown or HTML).
All that remains is for the extension to automatically generate a table of contents block at the top of every exported file, just like in books or in ChatGPT itself. Moreover, this can be implemented for all other LLMs supported by the extension (Gemini, Qwen, etc.), since they all split content into 'user question blocks' and 'LLM answer blocks.' Therefore, grabbing the user questions and generating a list at the top of the exported file shouldn't be hard, nor should linking the TOC items to the actual questions lower down in the chat. This would significantly improve navigation in long chats."
What would you like Inkpour to do? and Why is this useful?
"When a conversation in ChatGPT gets very long, it adds a kind of table of contents on the right next to the scrollbar (attached image), where it summarizes the user's question into a short heading. By hovering over this table of contents, you can quickly navigate through long chats.
Since ChatGPT already generates this, we could add the exact same table of contents at the top of the exported file, allowing users to click on these short questions to jump to them within the exported document. This means we already have all these headings inside the exported file (Markdown or HTML).
All that remains is for the extension to automatically generate a table of contents block at the top of every exported file, just like in books or in ChatGPT itself. Moreover, this can be implemented for all other LLMs supported by the extension (Gemini, Qwen, etc.), since they all split content into 'user question blocks' and 'LLM answer blocks.' Therefore, grabbing the user questions and generating a list at the top of the exported file shouldn't be hard, nor should linking the TOC items to the actual questions lower down in the chat. This would significantly improve navigation in long chats."