How It Works
Three steps.
One honest answer.
We don't tell you who to vote for. We show you what you already believe — then match it to what the candidates actually voted for and said.
Step 01
We dig up what candidates actually voted for — not just what they say.
For every race we cover, we pull official voting records, verified public statements, and campaign positions. No spin. No summaries written by their team. The actual record.
When what they did contradicts what they say, we go with what they did. Actions outrank words every time.
Source Hierarchy
Official government records + voting history
Candidate-submitted questionnaires
Verified media interviews + public statements
Third-party aggregators with documented sourcing
Step 02
We check every question for hidden spin before you ever see it.
Three AI models are involved — one writes the question, one scores it for bias, one picks the cleanest version. If a question has loaded language, slanted framing, or a built-in "right" answer, it gets rewritten or thrown out.
We use two different AI companies so no single model's blind spots take over. The whole audit trail is public. You can check our work yourself.
Pass 1
Generation
Claude generates 3–5 question variants from each candidate position using only behavioral language. No party names. No political labels.
Pass 2
Bias Scoring
GPT-4 scores each variant on four dimensions: loaded language, false equivalence, asymmetric framing, embedded cultural assumption. Scored blind — no knowledge of which model generated it.
Pass 3
Selection
The highest-scoring neutral variant is selected. All variants and scores are archived in the public audit trail.
Step 03
You answer honestly. We show you who matches.
Pick your state and your race. Answer the questions — not about the candidates, but about the actual issues. Do you support this policy? Is this something you care about? You don't need to know who anyone is.
When you're done, we show you which candidates in your race most closely match what you just said. No party labels in the results. Just the match score and why.
Take the awareness quiz →What You Get
Your belief profile
A clear picture of what you actually value, independent of party.
Candidate alignment scores
Which candidates match your positions and by how much.
Issue-by-issue breakdown
Where each candidate aligns and where they diverge.
Full methodology access
Every question, every bias score, every source — available to review.
No party. No endorsement. No algorithm built to make you afraid.
Read the full methodology →Ready to find your match?
The full platform launches before the 2026 election cycle.