Skip to main content

AetherGuard AI Documentation

AetherGuard is a zero-trust AI security gateway that sits between your application and your LLM provider. Every request passes through an 8-stage security pipeline — providing prompt injection prevention, PII redaction, toxicity detection, model attestation, and immutable audit provenance without changing your application code.

How it works​

Your App → AetherGuard Proxy → LLM Provider (OpenAI, Anthropic, etc.)
                  ↕
         8-Stage Security Pipeline

The proxy is fully OpenAI-compatible. Migrate by changing two values:

SettingBeforeAfter
Base URLhttps://api.openai.com/v1Your AetherGuard proxy URL
API Keysk-... (provider key)ag_... (AetherGuard key)

Key capabilities​

  • Prompt injection detection — blocks jailbreak attempts before they reach the model
  • PII/PHI redaction — detects and redacts personal data in prompts and responses
  • Toxicity filtering — blocks hate, abuse, and harmful content
  • Secrets detection — prevents API keys and credentials from leaking
  • Policy enforcement — per-tenant configurable rules (block, flag, redact, allow+log)
  • Cryptographic provenance — every request is signed and written to an immutable audit ledger
  • Model attestation — optional Nitro Enclave support for hardware-level trust
  • RAG security — chunk-level scanning, signature verification, and PII sanitization