You'd never open a suspicious attachment. Your AI opens them all.
One hidden line in an email, web page or document can hijack ChatGPT, Copilot or any AI assistant. Watchpost scans it before your AI reads it, and also catches AI scams and rogue agents. It tells you what to do, in plain language.
No sign‑up for your first scan. Everything is checked right on your device. Unlimited scans for $9.99/month, cancel anytime.
10 ways AI is used against you. Watchpost checks for all of them.
The free scanner checks emails, messages, web pages, scripts, app permissions and AI agent settings for signs of each attack. The full app will block them automatically, as they happen.
Device takeover
Fake "support" asking you to install remote‑control apps, download‑and‑run scripts, remote‑access malware.
Data theft
Attempts to send out your passwords, keys, contacts or files, and secrets you're about to paste into an AI.
Spyware & surveillance
Keyloggers, hidden camera, mic, screen or location access, and apps or extensions that watch everything you do.
Ransomware
Commands that delete your backups, code that encrypts your files, ransom notes.
Network attacks
Turning off your firewall or antivirus, redirecting your internet traffic, tampering with your router or Wi‑Fi.
AI phishing
Convincing scam messages, look‑alike links, fake login pages, "it's me, I lost my phone" family scams.
Privilege escalation
AI agents set to act without asking, and programs grabbing administrator powers they shouldn't have.
Prompt injection
Hidden or invisible instructions in emails, pages and documents that hijack your AI assistant.
Supply‑chain attacks
Fake look‑alike software packages, hidden install scripts, poisoned AI plugins and unsafe AI tool servers.
Autonomous agent attacks
Self‑spreading "AI worm" prompts, instructions passed between agents, agents acting without your OK.
Three steps. Under a second.
Paste or open it
An email, a text, a web page, a script, an app's permissions or your AI agent's settings.
Watchpost checks it
It looks for signs of all 10 attack types, including text you can't see, right on your device.
You get a clear answer
A risk score, what's wrong in plain language, and exactly what to do next.
Anyone who uses AI, or gets targeted by it.
Families
Check a suspicious text before your parents send gift cards to a "grandson" whose voice and message were faked by AI.
Small businesses
Catch AI‑written invoice scams and emails that try to hijack the AI tools your team relies on.
People using AI assistants
Scan documents, emails and pages before your assistant reads, summarizes or acts on them.
Developers & IT
Check AI agent settings, plugins, tool servers and install files for risky permissions and fake packages.
Your data stays with you.
The scanner runs entirely in your browser. Nothing you paste or open is uploaded, stored or seen by us. A security tool should never become a new place for your secrets to leak.
Try it free. Subscribe when you're ready.
Your first scan is free, no sign‑up needed. Watchpost Pro gives you unlimited scans, so you can check everything before your AI reads it. Subscribers also get first access to the full Watchpost app, which will stop attacks automatically as they happen.
Not ready yet? Get Watchpost news.
We'll tell you when the full app launches.
- Unlimited scans of emails, messages, pages and files
- Checks for all 10 AI attack types
- Everything stays on your device
- First access to the full Watchpost app
- Cancel anytime
Secure checkout by Stripe.