We read the boring documents
so you do not have to.
WhatItMeans.org exists for one reason: the information people need in order to vote, show up, or push back is almost always public — and almost never readable.
The mission
Democracy runs on documents nobody reads. A 104-page ballot booklet. A 300-page budget. A utility rate filing. A zoning amendment written in cross-references. These are usually honest and carefully made — and functionally useless to the person they are supposedly for.
We take one of those documents at a time and write down what it actually means. Short sentences. No jargon. No spin. If a measure is confusing, we say so. If the official title is misleading, we show you the gap. If we are not sure, we tell you that too.
Our one rule
Explain, never persuade. You will not find a voting recommendation anywhere on this site. Our job is to get you to the point where you can decide for yourself — and then get out of the way.
How we work
We start from the primary source
Every explanation traces to a named public document, linked so you can check us. Not a news summary, not a campaign flyer. If something is not in the source, it does not go in the explanation.
We use AI, and we show our work
We have two different AI models read the same source text separately, with the same instructions, neither able to see the other's answer. Then we publish both and point out where they diverged. That is not a gimmick. When two careful readings of one document come apart, that spot is almost always where the document is genuinely ambiguous — and that is exactly what you want flagged.
A person reviews everything before it publishes. AI drafts; humans decide. Pages about real people stay unpublished until a human has checked them.
We tell you who is behind it
Official summaries are legally required to be neutral, which usually means they leave out the most useful thing: who wrote this, who filed it, and who is spending money to convince you. That is public record. We put it on the page, with a Who is this? link on every name so you are never looking at a stranger.
We are honest about estimates
Some things we can state as fact: a vote tally, a dollar figure, a filing date. Other things are judgment — like our meter showing where a measure sits politically. Anything that is our read is labeled as our read, with the reasoning shown and a confidence level attached. We would rather say "we are not sure" than fake precision.
Where we are not neutral
We are non-partisan, but we are not indifferent. We think people should be able to understand what they are voting on, we think who pays for a campaign is relevant information, and we think a document written to be technically accurate but practically unreadable is a failure. Those are the biases. Now you know them.
When we get it wrong
We will. When you find it, tell us — use the request form and say what is wrong. We fix errors rather than quietly editing around them, and a substantive correction gets noted on the page.
What should we explain next?
The best projects come from people who are stuck on something specific.
Request a project See what we have done