# ArgusAI

> AI Firewall and agentic security gateway for LLM applications.

ArgusAI is an OpenAI-compatible security proxy that inspects chat completions, MCP tool calls, and agent workflows for prompt injection, PII leakage, and policy violations.

## Product

- AI Firewall for prompt injection and jailbreak defense
- PII detection and redaction on ingress and egress
- MCP tool call and output inspection (Growth plan+)
- Cross-tool taint tracking (Growth plan+)
- L1 pattern rules, L2 heuristic scoring, optional L3 LLM reasoning
- Audit logging and dashboard analytics

## Key pages

- Homepage: /
- Playground (try detection): /playground
- Documentation: /docs
- API reference: /api-reference
- Glossary: /glossary
- Security & trust: /security
- Research & benchmarks: /research

## Documentation

- Getting started: /docs/getting-started
- API overview: /docs/api
- MCP inspect API: /docs/mcp-inspect

## Solutions

- Prompt injection protection: /solutions/prompt-injection
- Data leakage prevention: /solutions/data-leakage-prevention
- MCP security: /solutions/mcp-security
- Agentic AI security: /solutions/agentic-ai-security

## Contact

- Website: https://argusai.io
- Full product summary: /llms-full.txt