Finally understand
what you’re reading
Select text in any PDF and ask AI to explain it.Highlight, comment, and organize your reading.
Features
Everything you need to
read with clarity
01 Topics
Chat across all your documents.
Group PDFs into topics and ask AI questions that span everything. Compare ideas, find connections, and synthesize across your entire library.
Machine Learning Research
Neural Networks.pdf
Attention Is All You Need.pdf
Scaling Laws.pdf
Compare how each paper defines “emergent abilities”
Across your 3 documents, the term is used differently. In Neural Networks.pdf...
02 Ask AI
Highlight text. Ask anything.
Select a passage you don't understand, ask AI to explain it, and get a clear answer in seconds. Powered by Claude, OpenAI, and Gemini.
What does this mean?
This explains how the model adjusts weights through backpropagation...
03 Verified
Every claim backed by your source.
Inline citations link every AI answer back to the exact page and passage. Low-confidence claims are flagged so you always know what's solid and what's not.
The transformer architecture uses self-attention[1] to weight token relationships, replacing recurrence entirely[2]
This approach may reduce training time by 40%[3]⚠ Low confidence — not well supported by source
Source [1] — p.4
“Self-attention allows the model to attend to all positions in the previous layer…”