PII Shield stops staff and AI agents leaking personal data into AI tools such as ChatGPT, Claude, Gemini and Copilot. It replaces names, ID numbers, email addresses, phone numbers, card and bank details with tokens like [PERSON_1] before anything is sent, then puts the real values back in the AI's reply, so the answer still reads normally and nobody has to change how they work.
One redaction engine covers every way a team uses AI: a browser extension for Chrome, Edge and Brave that masks the chat box before send; a drop-in gateway for the OpenAI and Anthropic APIs; an MCP server; a hosted connector that puts your ATS, CRM or helpdesk into Claude with records already tokenised; a Claude Code plugin; and format-preserving redaction of Word, Excel, PowerPoint and PDF files, including scanned documents. A Leak Audit shows what personal data has already been pasted into AI chats.
Prompts and personal data are never stored; logs hold counts only. Built in South Africa with POPIA and GDPR templates included. Public pricing, a 7-day free trial with no card, and self-hosting on the Enterprise plan.
