Hister stores only the text content of HTML/pdf pages. 1000 documents require around 80-100MB of storage and there is still plenty of room to optimize for storage space.

I'm using it for 6-7 months and my index size is below 1GB with almost 10k pages.

Also, a downside of the proxy approach: it does not handle properly JS based websites and cannot identify dynamic content changes. Our extension periodically checks if the browser tabs' content has been changed and automatically updates the index when change detected.