PromptSiren Free

Local-first claim-versus-reality checks for generated links, QR payloads, and install commands on AI assistant pages.

Stop AI chatbots from handing you phishing. AI assistants are great at summarising pages and finding links, but attackers have learned to abuse that. By hiding a few instructions on an ordinary web page, an attacker can make an AI's answer present phishing links, fake "your account is locked" alerts, malicious QR codes, and tracking images, all wrapped in the assistant's trusted interface. Because the link looks like it came from the AI, people trust it. PromptSiren Free is your local safety net. It reads what your AI assistant actually puts on screen and checks where links, QR codes, and commands really lead, then flags or blocks anything deceptive before you act on it. WHAT IT CATCHES • Lookalike and typosquatted domains (amaz0n, micros0ft-login, and similar tricks) • Homograph and punycode domain spoofing • Links where the claim ("the official login page") doesn't match the real destination • URL shorteners and redirectors that hide the true target • Malicious QR codes embedded in AI-generated images • Risky install commands (pipe-to-shell, fake or typosquatted packages) • Tracking "beacon" images that silently leak your IP, browser, and referrer PRIVATE BY DESIGN • Everything runs on your device — using Chrome's built-in on-device AI (Gemini Nano) plus a local detection engine. • No servers. No accounts. No sign-up. No analytics. No data ever leaves your computer. • It only runs on supported AI chatbot sites — it does not watch the rest of your browsing. FREE No subscription and nothing to pay. All the work happens locally, so there's no backend and no usage limit. WORKS WITH Works with most major AI Chatbots including ChatGPT. HOW TO USE IT Install it, open your favourite AI assistant, and you're protected. Suspicious links are disabled and labelled, malicious QR codes are removed, and risky content is flagged. Choose Block, Warn, Audit, or Human-review mode in options. Chrome's on-device AI adds deeper judgement when it's available; the local engine protects you either way. YOU STAY IN CONTROL • Warned links aren't a dead end: click the warning to see exactly where the link really goes (safely defanged), and click again if you decide to proceed. • Trust a site we flagged? Add its domain to your always-allowed list in options and it won't be flagged again — lookalikes of it still will be. • The popup shows your recent verdicts, including which domain was caught (never full URLs). A NOTE ON EXPECTATIONS PromptSiren Free raises the bar against AI-delivered phishing, but it's a safety net, not a guarantee. Keep verifying before you enter credentials or run a command. PromptSiren Free is the AI-chatbot companion to QRSiren (https://www.qrsiren.com), which protects against malicious QR codes. https://www.promptsiren.ai