WhoIsRight

When you ask an AI something like "Coffee is better than tea, right?", the answer will be biased — AI tends to agree with however the question is framed.

WhoIsRight fixes this:

1. Takes your question and automatically rewrites it from the opposite side
2. Sends both versions to multiple AI models in parallel — each rates both positions from 1 to 100%
3. The left and right branches know nothing about each other — fully independent
4. The average score across all responses reveals the real picture, free from framing bias

You can also enable "Persuasion" mode — the AI will argue for each side and reconsider its scores. After all rounds, you can continue the debate yourself!

Algorithm (step by step)

1. Parse & rewrite — The summarizer model reads your question, identifies both positions, then creates two biased versions — each favoring one side
2. Dual blind evaluation — Both questions go to every selected model at the same time. Left and right branches know nothing about each other
3. Initial score — Each model rates both positions 1–100%. Scores are averaged across all models and both framings — this cancels out the framing bias
4. Persuasion rounds (if enabled) — Each model builds the strongest arguments for both sides. If multiple models — arguments are merged by the summarizer. All models re-evaluate with the new arguments. Each round sees the entire debate history
5. Your turn — Add your own arguments. Every model re-evaluates the entire discussion including your input. Repeat as many times as you want
6. Final summary — All arguments are ranked by strength for both sides

Key principle: at every step, each AI sees the full history of the debate. The more rounds, the more refined the score.