WTF Clause

You clicked "I agree" without reading it. Everybody does. WTF Clause reads it for you. Open any website, click the icon, and in under a minute you get a plain English answer to one question: what does this company's own wording actually let them do with you and your data? WHAT YOU GET A risk level in green, yellow or red, so you know the answer before you read a word. The worst clauses, quoted word for word from their document, each with a short explanation of what that sentence permits. A share card you can post, carrying their quote, the address it came from, the date and second it was read, and a SHA-256 fingerprint of the exact text. HOW IT IS DIFFERENT Every finding is verbatim. If the exact wording is not in the document that was read, the finding is deleted rather than shown. The words are theirs, never ours. A second AI pass checks every finding before you see it. It reads the quote with the text around it and asks whether the claim really follows, whether the quote was taken in context, and whether the strength of the wording matches. Anything that fails is dropped. A clean policy comes back clean. This is not built to find something wrong every time. Wikipedia scores low. It is supposed to. It saves the full text it read, so a report can be checked against the exact document it was made from, and so you can tell when a company quietly changes the deal. NO KEY, NO ACCOUNT Install it and scan. There is nothing to sign up for and nothing to paste. By default the reading is done by the WTF Clause service. Your browser finds the policy and sends the service a fingerprint of the text; if somebody has already had that exact text read, the answer comes back from the cache in a moment. If not, the text is sent once, read with an AI model on our key, and kept for everyone who scans the same words after you. The service is sent the company's public policy text, the site's address and the tone you picked. It is not sent your name, your page title, your cookies, your history or anything you typed, and it keeps nothing about you. Prefer that nothing reaches us at all? Choose "My own API key" in the settings and add a key from OpenRouter, OpenAI or Anthropic. Then the document goes straight from your browser to the provider you picked, billed to you, and our service is never contacted. Your key is stored on your machine only and is never synced to your Google account. WHAT IT IS NOT Automated analysis, not legal advice. It reads what the document says. It cannot tell you what a court would decide. ABOUT THE FOUR WEBSITES CHROME WARNS ABOUT Chrome will tell you this extension can read and change your data on wmvvifhaffhwguzlhuqk.supabase.co, openrouter.ai, api.openai.com and api.anthropic.com. The first is the WTF Clause service. The other three are the AI providers you can choose instead. Those four are the only things this extension is allowed to contact: they are how your question reaches the reader and how the answer comes back. Chrome has no wording for "sends a request", so it warns about the whole address. No other site is in that list, and nothing else is in the box. Access to the pages you scan is asked for separately, in its own prompt, the first time a policy lives on a different address from the site you are on. You can say no and the extension keeps working. The code ships unminified. You can read all of it.