Free · runs on your own AI, not ours
Take our fact-checker to your AI.
Our fact-check instructions turn ChatGPT, Claude, Grok, DeepSeek or any AI you already use into a careful fact-checker. Copy the prompt, paste it into your AI, then paste the article or transcript right under it. We never run the check and never see what you check.
Instructions 3 steps — click to open
▶ Listen to these instructions
No account, no upload, nothing sent anywhere by us. The buttons just hand you our prompt to use in an AI you already have.
- Pick how deep — Quick, Standard or Deep.
- Hit "Copy the prompt."
- In your AI (ChatGPT, Claude, Grok, DeepSeek), paste the prompt, then paste the article, transcript or claim right under it, and send.
Best results: in our testing, ChatGPT, Claude and Grok did the best job. DeepSeek works with its Search button on. Some AIs turn web search off in private or temporary chats, so use a normal chat. For anything important, run it in two AIs and compare.
Keep in mind: the AI's answer is a well-sourced opinion, not a verdict from on high — read it the same way you'd read any second opinion, and check the sources it points you to.
2. Take the prompt
3. In your own AI
Paste the prompt, then paste the article, transcript or claim right under it, and send. (Downloaded it? Attach the file, paste your article in the message, and send.)
For the best result: use ChatGPT, Claude or Grok, and make sure web search is turned on — without it, an AI can only guess. Each AI checks a little differently, so for anything important, run it in two of them and compare.
Read the prompt first No hidden instructions — click to open
This is exactly what the button copies. Nothing hidden, nothing extra.
Works on anything. A news article, a YouTube transcript, a social post, a book excerpt, a claim a friend forwarded you — paste it under the prompt. That includes our own articles: paste one of our explainers under the prompt and see how it holds up under its own medicine.
What it is
A second opinion, on anything.
It hands your own trusted AI a clear job: check every factual claim, say what's solid, what's shaky, what's missing context, and what it couldn't verify — then give you a straight verdict. You read the answer; you decide.