AI Security Platform

The security layer for AI applications.

REST APIs to detect prompt injection, PII leaks, unsafe model output, and risky agent actions. One integration, structured verdicts, production-ready rate limits.

Core AI Security APIs

Purpose-built endpoints for each stage of your AI pipeline — input, output, and agent actions.

Prompt Injection Shield

Detect instruction override, role manipulation, and injection attempts in AI prompts.

PII & Secrets Detection

Find emails, phone numbers, API keys, private keys, and other sensitive data in text.

AI Output Safety

Classify AI-generated content for harmful, unsafe, or policy-violating output.

Agent Action Guard

Evaluate AI agent tool calls against policies — allow, review, or block actions.

Unified Guard

One request, multiple checks

Run prompt injection, PII detection, output safety, and agent action checks in a single call. Receive an aggregated decision with per-check breakdowns.

Explore Unified Guard
{
  "request_id": "req_01HXYZ",
  "api": "unified-guard",
  "decision": "block",
  "checks": [
    {
      "check": "prompt_injection",
      "verdict": "block",
      "findings": [{ "category": "instruction_override", "reason": "..." }]
    },
    {
      "check": "pii_secrets",
      "verdict": "block",
      "findings": [{ "category": "email", "reason": "Email detected" }]
    }
  ],
  "usage_units": 2,
  "processing_time_ms": 18
}

Built for developers

Structured JSON responses, consistent verdict semantics, and tooling that fits your existing stack.

REST + JSON

Standard HTTP POST endpoints. No SDK required — integrate from any language.

Structured verdicts

Every response includes verdict, risk level, findings, and reasons.

Usage tracking

Transparent usage units per check. Monitor consumption in your dashboard.

API key auth

Bearer token authentication. Keys managed in dashboard with rotation support.

How it works

1

Send input

POST text, model output, or agent action metadata to the relevant endpoint.

2

Evaluate

Detectors analyze content against injection, PII, safety, and policy rules.

3

Receive verdict

Get allow/review/block decisions with findings you can act on programmatically.

More APIs

Legacy validation, enrichment, and utility endpoints available on RapidAPI.

Simple pricing

Start free with 500 checks/month. Scale to Developer or Growth as traffic grows.

View pricing

Ship secure AI features this week

Create a free account, generate an API key, and run your first security check in under five minutes.

Get your API key