Read and annotate
Read PDFs with the agent beside you, highlight what matters, and keep your annotations in one list.
Open any document from your library to read it in the viewer. The document fills the main area and the right sidebar holds chat, details, and annotations, so you can read, ask, and verify without switching windows.
Move around
Use the toolbar to go page by page, type a page number to jump, or change the zoom. Reading progress is tracked automatically and shows in the library's Progress column.
Select text
Select a passage and a floating toolbar appears:
| Action | What it does |
|---|---|
| Chat | Sends the exact passage to the chat so your question is grounded in it |
| Highlight | Saves a color highlight as an annotation |
| Comment | Highlights the passage and attaches your note |
Use colors as categories, for example findings in one color and methods in another. Click an existing highlight to change or delete it. Overlapping highlights merge, and their comments are kept.
To understand a confusing sentence without starting a new conversation, use Explain on a selection in the PDF or in a chat reply.
Review annotations
The Annotations tab lists every highlight and comment in reading order. Click one to jump to that passage. You can also ask the agent about your highlights, for example "What do my highlights say about the study design?"
Check the details
The Details tab shows the file's metadata, an AI summary written from the full text, the author's abstract, and a formatted citation you can copy. It is a quick way to decide whether a paper is worth reading in full. Click any field to correct it.
Compare side by side
Open two documents in separate tabs to read them next to each other, or ask the agent to compare them. See Compare two papers.
If the chat gives vague answers about a document, check its Full text status in the library. Files marked Abstract only or Content unavailable have little text for the agent to read. See Add files.