Can DEVONagent search and summarize papers in a DEVONthink database without using DT Server?

Is that possible? I am just exploring the feasibility of having these two powerful apps connected. One could add 100s of PDFs in a DT database and use DA to search keywords, digest the relevant PDFs, and summarize their connections.

Not without Server, so far as I’m aware. You’ve probably seen that the promisingly-named “DEVONthink databases” plugin in the Browsers section of the Plugins only scans bookmarks in your open databases, whereas the “DEVONthink Server” plugin does scan all open databases via the local network.

It is a brilliant function, though – almost worth the hefty cost of Server in itself. Many’s the time I’ve used DA to scour the web for something, only to find I had forgotten or unnoticed matches in my own local databases.

1 Like

Thanks. That is sad. @cgrunenberg, maybe this linkage is something to be added to your list :).

I’m not sure what kind of summarization you’re referring to but @NickLowe is on the right track. If you’re running the Server edition, the DEVONthink Server plugin in DEVONagent Pro can locate PDFs in your open DEVONthink databases. However, the DEVONagent account must have at read access permissions in the Server settings for the database you want to search.

The same type of summary that DevonAgent does when find different webpages.

Are you referring to to Digest ?

Yes. Imagine Devoagent digesting PDFs found in DT. Game changer for researchers!

Well, I wouldn’t get too excited. At least the way a search through Server does it, there’s no way of restricting the search scope to particular groups, and you have to close all open databases except the ones you’re searching, while as always with DA the quality and quantity of summary varies unpredictably from hit to hit, and some databases report no results at all, for reasons I’ve never figured out. (It’s not to do with the DA permissions set for that database in DT’s app preferences.) And you only get one summary per document, as opposed to DT’s deep-drill view in the inspector of all the matches within the document with a few lines of context for each.

I agree that something like DA’s summarisation and digest features would be a fantastic enhancement to DT, but I suspect it would need to be implemented in-app rather than in DA, which is very specifically designed around web searches and whose interaction with DT depends on essentially fooling it into treating your open DT databases as a crawlable website. The value of DA searches of DT Server for me is less in the summaries (handy though those are in reminding me of what might be relevant in a file I wouldn’t have thought to look in) than in alerting me to documents already in my database that I’d completely forgotten about – quite often, embarrassingly, documents authored by me that already answer my own question…