Why I built Refract
A note from the person who built this
I was educated in the UK, and one thing from studying History in the 1990s has stuck with me ever since: whenever you're handed a piece of information, ask three things about it. Who's the source? What's their intention? Who's backing them?
Question the motive — not cynically, just habitually. It's a small discipline, but it changes how you read almost everything.
I've tried to pass that on to my own children. Not to distrust everything, but to not accept things purely at face value — even when it comes from someone they trust, a teacher, or me. Ask where it's from. Ask why it's being said this way. Ask who benefits if you believe it.
That instinct feels more necessary now than it ever did in a classroom. Social media and a near-infinite number of information sources have made it genuinely hard to know what to believe. Everyone's competing for attention, and the loudest framing usually wins over the most accurate one.
Refract isn't trying to tell you what to think, or steer you toward any particular conclusion. That's deliberate. The goal is to hand you the same kind of questions I learned to ask thirty years ago — about bias, funding, purpose, and whether the facts actually hold up — and let you draw your own conclusions with better information than a single headline gives you.
I'll be honest about the limits of this too. It's built on AI, and AI gets things wrong sometimes. It isn't a perfect, infallible verdict machine, and I don't present it as one. But even an imperfect second opinion — checked against three different AI models rather than trusting just one — is more than most of us bring to our news feed on a normal day. For me, that's genuinely valuable: not a replacement for thinking, but a tool to help you do it a bit better.
That's really the whole idea. Question the source. Question the motive. Then decide for yourself.