Is this actually true — ask three bots

Ask one chatbot about something you read and you get one answer. To check it you open another tab, retype the same sentence, and compare the two by eye. WhipWhip folds that round trip into one run.

How the run goes

1. You paste the claim once. A link, or something you were told — as it is.
2. Three services are asked in turn. Perplexity, ChatGPT and Claude each get the same claim in a new tab of their own. Perplexity is asked to find current sources; the other two are asked to answer from what they know, without searching — so a sourced answer stands next to what the models already knew.
3. The answers land on one screen. You read them next to each other.
4. Disagreement is the signal. When all three point the same way it is usually safe; when one goes its own way, that is where to dig. You are reading where they disagree, not how confident any one of them sounds.

Why one source is hard to trust

A chatbot leans toward filling a gap plausibly rather than saying it does not know, and that failure is hard to see from inside a single model — asked to review its own answer, it usually passes it. Give the same question to three different models and the failure shows up as disagreement. Handing one bot's answer to another for review is the same idea, and WhipWhip does the handing as a step instead of by hand.

What you need

A desktop Chromium browser — Chrome or Edge — and a Claude sign-in. Perplexity and ChatGPT work signed out. The procedure goes through all three, so if you are signed out of Claude it stops at that turn. WhipWhip is free and asks for no sign-up. WhipWhip never sends your conversations to us.

How to take it

This workflow is on the board. Copy the one string in the post and paste it into 📥 Import in the extension panel's Select whipset dropdown — a preview comes up first.