NNScholar

NNScholar is an AI research workspace covering the whole academic workflow: start a literature review from a topic, DOI, PMID, arXiv link or research question; search scholarly sources and import references; use AI to screen papers; chat with PDFs with translation and annotation inside the paper's context; trace every argument back to its source; and draft with research agents and academic skills — papers, notes, conversations, evidence, tasks and drafts all live in one project. Runs as a desktop workspace with a WebUse browser mode.

The evidence-tracking design is what makes it research-grade rather than chatbot-grade: claims you can trace to sources survive peer review; vibes don't.

Who it's for: grad students and postdocs managing literature reviews, researchers writing survey papers, and evidence-synthesis teams. Free entry with $7-ish tiers for volume. The decisive test is citation accuracy — have it verify five references you already know cold; scholarly AI that hallucinates a DOI is worse than no assistant, and one that catches itself is worth paying for.