Yumi: Research Without the Tab Apocalypse
I've spent too many hours with 47 tabs open: 12 research papers, 8 ChatGPT threads, 3 note-taking windows, and a Google Doc that's somehow still blank. Yumi's pitch is exactly this problem — put the AI, the notes, the sources, and the draft in one window so you stop context-switching yourself into exhaustion.
The execution is better than I expected from a new product. The three-panel layout (notes on left, web viewer center, AI chat right) feels natural within 10 minutes. You outline your structure in notes, browse sources in the center panel, and ask the AI questions on the right. When you find a useful stat or quote, one click saves it as a citation. When you're ready to write, the AI generates a draft that references your saved citations.
The citation system is the feature that actually matters. Most AI research tools either hallucinate sources or require you to paste URLs manually. Yumi's built-in browser means the source is literally open in the same window — the AI can see what you're reading and ground its responses in that page. I tested this with a dense academic paper on transformer architectures and the AI correctly pulled specific numbers from section 3.2 without fabricating anything. That's the bar. Yumi clears it.
Where it needs work: it's desktop-only right now, so mobile research sessions are out. The UI has some rough edges — I hit a scroll conflict between the notes panel and the web viewer twice. And the user base is small, which means fewer community templates and fewer people stress-testing edge cases. This isn't a dealbreaker for an early tool, but if you rely on community-shared workflows, you'll find the ecosystem thin.
I'd recommend Yumi for anyone who does research-heavy writing — students, analysts, content strategists, technical writers. The tab-reduction alone is worth the download. If your work is quick-turnaround social media posts where you don't need sources, this is overkill. Wait for the web version if desktop-only bothers you.
