Media
4 items
Description
Protect names, financials and credentials before you send them to AI. AI Safety Assistant is a spell-checker for confidential information. On ChatGPT, Claude, Gemini and Perplexity it watches the message you're composing and, before you send it, helps you catch and mask sensitive details. - Secrets never leave your browser. API keys, passwords, tokens and similar credentials are detected and masked locally, on your device, they are never transmitted anywhere. - Names & financial details. Only after removing secrets locally, the remaining text is checked by VaultLM's EU service to help find names and organisations (measured ~87% recall for names, human review still matters). One click replaces them with neutral placeholders like [NAME_01]. - Reversible for you, not for the AI. The mapping from placeholder back to the original stays with you (a "safe key"), so you can restore identities in the AI's reply on demand; the AI only ever sees the masked text. - Private by design. Processing happens on EU infrastructure. No advertising, no third-party trackers, no selling of data. This is a safety net, not a guarantee: always review a message before sending.