How it works
From your first paper to your final submission.
Step 1
Create a project
"One project. One knowledge base. Everything scoped to what you're working on."
Create a focused or exploration project. Choose your target venue. Your entire library, highlights, and writing stay isolated per project — no AI cross-contamination between research topics.
Step 2
Add and read your papers
"Upload a PDF. Read it on any device. Highlight what matters."
Drag, drop, or import via DOI or arXiv URL. Open on desktop or mobile. Select text to highlight — choose a category color. Every highlight indexes into your project's knowledge base instantly. Paper summaries are generated automatically in the background.
Step 3
Annotate and build your knowledge base
"You read. The workspace remembers. Nothing interrupts."
As you highlight, your knowledge base grows. Color-code by category — Evidence, Method, Result, Gap. Your annotations become the foundation for every AI check: logic review, novelty scoring, and writing assistance all draw exclusively from what you've read and marked.
Step 4
Write in LaTeX or plain document
"Draft in the way that makes sense to you. Export to the format the journal wants."
Switch between rich text and LaTeX at any time. The same document. As you write, citation autocomplete pulls from your library. Unsupported claims surface as quiet inline flags — not popups, not interruptions.
Step 5
Compile, check, and submit
"Run one final pass before it leaves your hands."
Compile your LaTeX against your target venue template. Run a full logic and novelty check across your document. Download the AI usage report if your institution requires transparency documentation.