Store history of summarizer requests

fredmb

I'm typically using Universal Summarizer for long-ish PDF documents. It would be nice to have a page with previous requests so I could quickly consult my previous summaries and "key moments". I'm currently copying the results and storing them locally.

I did notice that repeated requests didn't consume additional tokens since the summaries were already cached. Can I rely on this to be the case over time or is it a more temporary cache? It would at any rate be nice to have an interface for previous requests, since otherwise I'd need to keep track of the URLs and the metadata of the documents in order to re-submit them for summarizing.

I realize that for many use cases it's not necessary or even desirable to store the summaries permanently, so it ought to be possible to opt-in/-out and delete history.

Update: I learned via Discord that cache is regional (i.e. not shared globally) and lasts for 24 hours, although this may change in the future.

Vlad

fredmb Can you explain where would this history be? At Kagi we are constantly fighting the need to respect privacy and improve user experience, and by default we do not log or store anything.

How would you expect this feature to work? Can you describe your use case a bit more?

Spirits-

The itemized summaries that the summarizer produces would take up very little storage space, as the original document is not required anymore, and could be stored in a browseable history so you can reference them again.

Vlad

Spirits- We do cache summaries, up to a day.

firstohit

I have the same request. I would like to be able to access all previous history for research purposes. my motivation is that I don't want to use up all my tokens by requesting summaries I have requested before. At least 30 days or so.