Secure AI Prompts Before They Leave Your Device

On-device prompt analysis for iOS and macOS. ML-powered detection of injection attacks, jailbreaks, PII exposure, and secrets — before prompts leave your device.

Free tier available. No account required. All analysis on-device — including ML inference.

ML-Powered DetectionLocal-FirstMulti-ProviderCost EstimationmacOS & iOS
PromptGuard App Icon

Why PromptGuard?

Every time you paste information into ChatGPT, Claude, or any AI assistant, you're making a split-second decision about what's safe to share—and most people get it wrong. PromptGuard puts a security expert in your pocket, analyzing your prompts locally on your device before they ever leave your device. It catches the email addresses you forgot to remove, the API keys hidden in that code snippet, and the subtle prompt injection and jailbreak patterns that a neural ML classifier catches but simple rules would miss. For professionals who rely on AI daily but handle sensitive information—developers, lawyers, healthcare workers, executives— PromptGuard is the safety net that lets you work faster without the constant anxiety of wondering what you just accidentally exposed.

Security features for every prompt

PromptGuard analyzes your prompts locally before they leave your device, protecting you from security risks and helping you understand API costs.

Security Detection

ML-powered detection of injection attacks, jailbreaks, PII exposure, and secrets — all analyzed on-device using Meta's Prompt Guard neural classifier.

Local-First Privacy

All analysis runs on-device — including the neural ML model. Your prompts are never sent to external servers for security scanning.

Multi-Provider Support

Works with OpenAI, Anthropic, Ollama, and LM Studio. Secure prompts for any AI provider you use.

Guided Generation

Step-by-step secure prompt building with templates. Create effective prompts that follow security best practices.

Cost Estimation

Know your API costs before execution. See token counts and estimated spend for any prompt.

App Lock

Protect your prompts with Face ID or Touch ID. Biometric security keeps your AI conversations private.

Share Extension

Analyze prompts from any app via the iOS Share Sheet. Quickly check text for security issues without leaving your current app.

Home Screen Widgets

Security status at a glance with home screen widgets. Monitor your prompt safety score and recent analysis results.

iCloud Sync

Sync prompt history across devices with iCloud. Your analysis history follows you seamlessly between iPhone, iPad, and Mac.

macOS App

Native macOS app with full feature parity. Analyze prompts on your Mac with the same powerful security engine.

On-device ML detection powered by Meta Prompt Guard (DeBERTa)

Simple, transparent pricing

Start free with essential security features. Upgrade to Pro for advanced capabilities and guided prompt generation.

Free

$0

Get started with essential prompt security

  • ML-powered security analysis
  • 1 local server connection
  • 30-day prompt history
  • Neural injection & jailbreak detection

Pro

$9.99/mo

Advanced features for power users — 7-day free trial

  • Guided prompt generation
  • Cloud provider support
  • Unlimited local servers
  • iCloud sync
  • Prompt export & sync
  • App lock (Face ID / Touch ID)

Family

$14.99/mo

Share Pro with your household

  • Everything in Pro
  • Share with up to 5 family members
  • Family sharing via Apple
  • Separate prompt histories

Stop Hoping Your Prompts Are Safe.
Start Knowing.

Download PromptGuard and analyze your AI prompts for security risks before they leave your device.