New Updates to Scite Collections

Hey there.

I'm Drew Barontini, VP of Product for Scite, and I want to walk through some updates we've made to collections.

So you can see I have a bunch of different collections here saved in my private list.

I also could look at anything shared with my team through my organization account.

But one of the changes we've made here is to now allow you to create a blank collection.

So you only need a name, and that's the only required field to create a collection.

You can optionally include a description.

You could include the DOIs if you want, but you no longer have to have at least one doi.

You can just create a named collection here.

So you could do something like bookmarks or save papers, which is a great way to create just a blank collection that you could add papers to if you don't want something that's specific to a certain topic.

All right, so I'm going to jump into a individual collection here.

So I have this Optimal Exercise collection.

It's got 20 papers in in it, and I can manually remove these papers individually.

I can collect or select multiple of them and remove them because I've manually added these papers from assistant queries and searches and going to individual papers as I've used site.

Now if I go back to the, collections list and we look at this hit and blood pressure collection, you can see there's a lightning bolt icon that indicates this is a collection from a site saved search that will automatically update.

So the number of papers will change as new papers are added or new citation statements are added.

It will update over time, which is really powerful.

But what you can see here is inside of this individual collection.

Before, there wasn't a way to manually change anything inside of a collection that is a saved search collection.

So if I go to edit collection here, you can see it actually won't let me change those manual DOIs that I entered.

And this will update over time.

But now what we've done is we've split it between the list of all of the papers that are both manually selected and automatically saved from search results.

Or you could just look at anything you've manually selected.

So I've actually manually added this one paper.

I've only have one that's manually added.

And then if I look at the search results list, I have over 4,200 that have been added to, this particular list.

And you can look at these now and you can manage the full set.

New papers will come in and will be added, but any of the manual changes that you make will be reflected back in this collection.

To give you a lot more power with how you use it, We've also brought collections into the MCP experience.

So if you're using Claude or ChatGPT or Copilot or Gemini or anything that works with MCPS, you can now save papers that you find back into a collection.

You can create collections, you can edit them, you can delete them, you can do whatever you want to with those collections.

And this is really powerful because it becomes like this context container that you can carry through to different AI tools you're using.

You could work with Claude code and create a collection and save some papers there and then open up a separate session that is inside of ChatGPT or Codex and you could pull that same collection and make changes to it.

So that's what I'm going to show here is we're going to say, take my look at my optimal exercise site collection and give me a list of the marked citations in a simple table for the papers.

So if I just run this quickly, it's going to go and look at that specific collection I have called optimal exercise.

You can see it's searching the collections, it found the tools it can use for finding collections and pulling information from them.

And it's going to run a collection request to get all of the data.

And then it will also use another tool we have to search literature.

So it's going to get that collection.

It's going to say, okay, cool, here are the collection collection, here are the DOIs.

And then it's going to go run search literature, which allows it to build a very rich network of data that you can pull for all the metadata you want to pull back for the tallies of supporting, contrasting and mentioning.

And then this will come back in a simple table that will show us our response.

All right, so you can see now we're coming back with the 20 papers that were in that optimal exercise collection.

you can see that it's going to pull back the paper title, it's going to pull back the total number, which is the combined total number of citations, the supporting, the contrasting and the mentioning, and then it will give an example of any of that information in a simple table here so this all comes back, and it is the same data set you can operate on to do anything you want.

You could go and search and find more related papers, and you could add them to this collection.

you could modify the collections, you can do whatever you need to.

And as a reminder, you're always going to have this collection set up with alerts.

So if you want to get an email alert for when new citations are made to this particular collection, you can set that up, you can get that feedback, and we're going to continue to improve a lot of things around the feed and updates now that you start storing and curating the papers you're trying to find that are verified papers, and real DOIs that have all of this classification and quality context that you can carry with you to all of the different LLMs that you use.

If you have any feedback, please let us know.