Why we built TrueFrame
Last year, I watched two close friends argue for an hour about a news story. They agreed on the facts. The disagreement was about what those facts meant. One had read coverage from CNN and The Guardian. The other from Fox News and the Wall Street Journal. Same event, completely different narratives.
Neither was lying. Neither was stupid. They were just seeing the world through different editorial lenses — and they didn't even realize it.
The Invisible Filter
Most people get their news from 2-3 sources. They trust those sources. And those sources have editorial perspectives. This isn't a conspiracy. It's how journalism works. Editors choose what to cover, which facts to emphasize, which experts to quote, which headlines to write. These choices compound into a worldview.
The result: millions of people see fundamentally different versions of the same events. Not because the facts are different, but because the framing is.
What If You Could See Everything?
That's the core idea behind TrueFrame. What if, instead of choosing the "right" source, you could see ALL the sources side by side? What if every story showed you the full spectrum of coverage — from left to right — with transparent bias ratings?
Not to tell you which side is correct. But to show you how the same facts become different stories depending on who's telling them.
Why Now?
Two things made TrueFrame possible in 2026 that weren't feasible five years ago:
- AI can rate at scale. We rate over 10,000 sources for bias, factuality, and ownership. Doing this manually would require hundreds of analysts. GPT-4o can apply consistent criteria across every source and provide written rationales explaining each rating.
- Embeddings enable clustering. We group articles about the same event automatically using vector similarity. When 47 outlets cover the same Fed announcement, we show you one story with 47 perspectives, not 47 separate items in your feed.
What's Next
We're in private beta. Everything works, but we know it's not perfect. Our ratings will have errors. Our clustering will sometimes merge things that shouldn't be merged or split things that should be together.
That's why we built the override feature from day one. If you disagree with a rating, you can set your own. Your override only affects your experience. We think that's the right approach: provide a baseline with transparent methodology, then let you customize.
If you're reading this, you probably care about getting a full picture of the news. That's exactly who we built this for. Welcome.
— The TrueFrame Team