Skip to main content
PromptGuard scans every LLM request and response for security threats — prompt injection, jailbreaks, PII, data exfiltration, toxicity, and more — starting with a deterministic fast path that runs in under 10 ms. By default a request the fast path finds clean continues to the ML classifiers, so most benign requests take longer (latency budgets). Add one line of code to protect your entire application.

Quickstart

Get protected in 5 minutes

Why PromptGuard?

What it protects and why — no code

API Reference

REST API with interactive playground

MCP Server

Connect to Cursor, Claude, VS Code

​
Two products, one platform

Both run on the same detection engine and policies, and both can be fully self-hosted — including air-gapped for regulated environments.

​
How it works

Three ways to integrate:

​
What we detect

Prompt Injection and Jailbreaks

ML ensemble plus LLM-powered analysis across 7 attack categories, including multi-turn escalation.

PII and Secrets

46 entity types with checksum validation. API keys, tokens, and credentials with entropy analysis.

Content Safety

Toxicity, multi-turn intent drift, streaming output guardrails, and MCP tool security.

AI Agent Traps

21 attack vectors from DeepMind’s framework: steganography, RAG poisoning, sub-agent spawning, and more.
Last modified on October 31, 2026