SOLUTIONS / AGENT BUILDERS

Protect your agents

Drop-in prompt injection protection for MCP and A2A agents. Layered latency — local deterministic rules <0.1 ms, Galvanize-60M CPU inference at 11.52 ms p50 (sub-15ms) — means your agents stay fast while staying safe.

View pricing Read the docs

Technical specifications

sub-15ms CPU

Latency

11.52 ms p50 CPU · local rules <0.1 ms

1.00%

Tool False Positive Rate

certified across 9/9 industrial gates

MCP + A2A

Protocols

native support

90 days

Evidence Retention

audit trail included

Integration

zn is an HTTP API you can call from anywhere — any language, any agent framework. A Python client is bundled with the source release. Add zn to your agent in minutes:

bash

curl -X POST https://api.usezn.com/v30/analyze \
  -H "Authorization: Bearer $ZN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"input": "the tool-call input your agent is about to execute"}'

# verdict "block" -> reject the input before the tool call
# verdict "allow" -> safe to proceed

Features

Real-time prompt injection detection

Deterministic, versioned rules engine

Evidence vault with SHA-256 hashing

Structured threat classification

Zero-config MCP integration

A2A protocol support

Usage alert emails

Rate limiting and quotas

How it works

1. Intercept

Every tool call passes through zn before reaching your agent. We analyze the input in real-time.

2. Analyze

Our detection engine identifies prompt injection patterns, instruction overrides, and data exfiltration attempts.

3. Protect

Threats are blocked, evidence is logged, and your agent continues operating safely.

Ready to protect your agents?

Start free with a 7-day trial. No credit card required.

View pricing