How it works

The anatomy of a Prism paper

Every explainer is built from the same parts, so once you've read one, you know your way around them all.

01

Ingest

We start from the original paper — its text, math, figures, and results.

02

Refract & fact-check

We rewrite it as a plain-language narrative, then verify every claim against the source before it ships.

03

Bring it alive

We design custom interactive widgets so the hardest concepts become something you can play with.

What's inside every paper

TL;DR + key numbers

Every paper opens with a plain-language summary and the handful of numbers that actually matter — so you get the gist in 30 seconds.

A narrative, not a wall of text

We rebuild the paper as an ordered story: the problem, the core idea, the mechanism, the results — each in plain language with a real-world analogy.

Interactive widgets

Where a concept is easier to feel than to read, we drop in a widget you can hover, drag, and step through until it clicks.

Two levels at once

Read the plain-language version, or dig into the accurate technical detail. You choose your depth, paragraph by paragraph.

Glossary on tap

Every bit of jargon is defined in one place, in words anyone can understand — no prior knowledge assumed.

Key takeaways

We close every paper with the handful of ideas worth remembering, so it sticks after you leave.

See it for real on our first paper — the one that started modern AI.