Folder chat lets you ask questions across an entire collection of documents at once. Instead of consulting papers one by one, you get synthesized answers drawn from everything in the folder, with citations to every source.
When to use folder chat
Folder chat is the right tool when your question spans multiple documents:
Synthesizing findings across a literature review collection
Comparing methodologies or results between studies
Finding contradictions or agreements across papers on a topic
Extracting specific data points from many documents (sample sizes, key metrics, dates)
If you need focused, deep answers about a single paper, use the file chat in the document viewer's Ask tab instead.
How to start a folder chat
On the Home page, click Chat with folder in the Quick Actions section.
Select the folder you want to chat with.
A new chat opens scoped to that folder. You'll see the folder name shown as context.
You can also navigate to a folder in the Library sidebar and start a chat from there.
How the AI searches your folder
When you ask a question, Anara searches across every document in the folder for relevant passages. Subfolders are included automatically — nested structures are fully covered. Every claim in the response has an inline citation you can click to go to the source. Files in your library open in the document viewer and scroll to the exact passage.
Good to know: Anara searches across folders of any size. There's no limit on how many documents a folder can contain — the AI handles large collections just as well as small ones.
Using @mentions to narrow scope
You can focus the AI on specific documents within the folder using @mentions. Type @ in the chat input and select a file from the Files list. The AI will prioritize that document while still having access to the rest of the folder.
Tips for best results
Be specific. "What sample sizes were used across these studies?" works better than "Tell me about these papers."
Try different models. Click the model selector next to the chat input to switch models. More capable models (available on Pro and Max plans) handle complex synthesis better.
Use @Plan for multi-step tasks. Type @Plan to ask the AI to outline an approach before diving in — useful for systematic literature reviews.
If something isn't working
The AI seems to be missing some documents. Check that the files have finished processing. In the library table, the "Full text" column should show "Available" for a document to be fully searchable.
Answers aren't drawing from the folder. Make sure library search isn't disabled. Check the toggle buttons below the chat input.