Open Source LLM Security

Protect Your AI with
Enterprise-Grade Security

High-performance guards against prompt injection, jailbreaks, and adversarial attacks. Built in Rust for speed. Designed for production.

Built for Security Teams

Multi-Layer Defense

Combine pattern matching, ML classification, and semantic analysis for comprehensive protection.

Blazing Fast

Built in Rust with zero-copy parsing. Process thousands of requests per second with sub-millisecond latency.

Compliance Ready

Generate reports for NIST AI RMF, EU AI Act, and SOC 2. Built-in audit logging and telemetry.

Simple Integration

Add protection to your LLM pipeline in minutes.

use oxideshield_guard::{MultiLayerGuard, GuardConfig};

let guard = MultiLayerGuard::new(GuardConfig::default());

let result = guard.check("user input here").await?;

if result.is_blocked() {
    // Handle blocked input
    println!("Blocked: {}", result.reason());
}

Simple, Transparent Pricing

Start free with our Community edition. Upgrade as your security needs grow.

Monthly
Annual 2 months free

Community

For individuals, evaluation, and open source

£0 /month
Free forever
Get Started
  • PatternGuard - Aho-Corasick + regex
  • LengthGuard - input validation
  • EncodingGuard - Unicode attack detection
  • PerplexityGuard - adversarial suffixes
  • PIIGuard - detection & redaction
  • ToxicityGuard - content moderation
  • AuthoritarianUseGuard - responsible AI
  • CLI tool & Python bindings
  • Community support (GitHub)
  • Resource Limiter
  • ML-powered guards

Enterprise

For large organisations and regulated industries

POE
Price on Enquiry
Contact Sales
  • Everything in Professional, plus:
  • Dashboard - real-time monitoring UI
  • REST API for integrations
  • Webhook Alerts (Slack, Discord)
  • Rate Limiting per client
  • Streaming Guards (SSE)
  • Custom Guards development
  • Private/self-hosted models
  • SSO/SAML integration
  • Priority SLA-backed support
  • Dedicated CSM