Verho

Finds personal data in your prompts to Claude or ChatGPT, swaps it for placeholders when you ask, and puts it back in the answer.

Verho finds the personal data in what you are about to send to Claude or ChatGPT, and replaces it with placeholders when you ask. When the answer comes back, Verho puts the real values back into it, in your own browser. Keep using the AI tools you already know. Verho shows what is sensitive; you decide what is protected, and you send the message yourself. HOW IT WORKS 1. Write or paste your text into Claude or ChatGPT as usual. 2. Verho highlights the personal data it finds: names, personal identity codes, addresses, email addresses, phone numbers, bank account numbers, salaries, health details, passwords and more. 3. Add or remove items by hand. Nothing in your text changes until you press Protect. 4. Protect replaces each value with a placeholder such as [PERSON_1]. You then send the protected message yourself. 5. When the answer arrives with the same placeholders, Verho puts the original values back into it on your screen. WHERE YOUR TEXT GOES • To find the personal data, the text in the message box is checked on the Verho server in Helsinki, Finland, inside the EU. The server keeps none of it: not in a database, not in its logs. • Which placeholder stands for which value is kept only in your browser, encrypted, in the extension's own storage. • Verho sends nothing to Claude or ChatGPT. You send your message yourself. • We know who you are from your account (email only), but not what you did: nothing that describes your use points to you. • The only exception is feedback you send from the extension, if you choose to attach the protected conversation. It is deleted after 90 days. ADDITIONAL INFORMATION • You need a free Verho account. Sign in with Microsoft, Google or your email address; your first sign-in creates the account. • Light, for individuals, is free. Pro, for firms and power users, is coming soon. • Verho works on claude.ai and chatgpt.com, in Chrome and Edge on a computer. • It recognises personal data best in English, Finnish, Swedish, Norwegian, Danish, German, French, Spanish, Italian, Dutch and Polish. • Detection is AI-assisted and can miss something, or mark something that is not sensitive. It is user's responsibility to look the prompt over before sending. Privacy notice: https://verho.ai/privacy/ Terms of service: https://verho.ai/terms/ Verho is made by Tekoälytehdas Oy in Helsinki, Finland. It is not affiliated with Anthropic or OpenAI. Claude is a trademark of Anthropic, and ChatGPT a trademark of OpenAI.