Read less.
Understand more.
Signal is an AI-powered research assistant that teaches you to read like a researcher — not just skim like a student. Paste in any paper or article and Signal will break it down, highlight what matters, and push you to form your own position before showing you the answers.
Most students read for coverage, not comprehension.
You've read the paper. You can summarize it. But can you identify its central argument in one sentence? Can you name its two biggest weaknesses? Can you say whether the evidence actually supports the conclusion?
These are the skills that separate a student who reads papers from a researcher who understands them. Signal is built to close that gap — not by giving you answers, but by teaching you what questions to ask.
Three steps to real understanding.
Paste raw text or upload a PDF — a research paper, news article, policy report, essay, or anything else worth reading carefully. Signal accepts up to ~13,000 characters and strips away formatting noise to focus on the writing itself.
Before Signal shows you anything, it asks: what do you think the argument is? What's the biggest weakness? Writing your take — even briefly, even uncertainly — forces active engagement. It's the difference between reading and thinking. You can skip this step, but your feedback will be sharper if you don't.
Signal generates a structured analysis of the document and color-codes the original text with key passages. If you submitted a take, you'll also get a comparison: where you were right, where you missed something, and what to look for next time.
Six things every skilled reader looks for.
Signal highlights 12–20 key passages directly on the original text, color-coded by type. Over time, you internalize the pattern — and start seeing it without the colors.
Eight lenses. One document.
Signal's analysis is structured around the questions a careful reader — or a good peer reviewer — would ask. Each section is designed to surface something the text alone doesn't make obvious.
Built to teach, not to replace.
Signal is built on a simple belief: the bottleneck in research isn't access to papers — it's knowing how to read them. Most people skim for conclusions. Signal teaches you to read for structure: how an argument is built, what evidence actually supports it, where the cracks are. The highlights aren't decoration — they're a map of how rigorous thinkers move through a text. The "Your Take" step isn't optional polish. It's the whole point. Forming a position before seeing analysis is how you build the mental muscle that makes you a better reader over time, not just on this paper.
A chatbot answers your questions. Signal questions your answers.
Ask a general AI to explain a paper and it will — clearly, confidently, and instantly. That's exactly the problem. When someone else does the thinking for you, you get a summary you'll forget by tomorrow. You don't build the mental model. You don't catch what the author assumed. You don't know what you would have missed.
Signal is built around a different premise: that the act of forming your own position — before seeing any analysis — is where the real learning happens. It's not a faster way to get answers. It's a structured way to develop the reading instincts that make you not need the answers in the first place.
General AI is optimized for convenience. Signal is optimized for comprehension. The difference shows up in small ways: Signal asks what you think before it tells you anything. It highlights the text you actually need to read, not just a clean rewrite of it. It tells you where your take was wrong, not just where the paper was right. None of that is an accident — it's what separates a tool that replaces thinking from one that builds it.
Built by students, for students.
Signal is an independent project — no lab, no funding, just a genuine belief that reading carefully is a skill worth building.
Ready to read differently?
Drop in any paper and see what you've been missing. It takes about thirty seconds.
Open Signal