Skip to content

Zentinelle

Governance, Risk, and Compliance for AI agents.

Zentinelle sits between your AI agents and their LLM providers, enforcing policies, scanning content, logging activity, and giving you a real-time GRC portal — without modifying agent code.

v1.2.0 — Production Ready

24 policy evaluators · 24+ LLM providers · Multi-cloud Terraform · 5 SDK languages · 3 framework plugins · MIT Licensed

How it works

Your Agent  ──►  Zentinelle  ──►  LLM Provider
               (policy check)    (Anthropic, OpenAI, Google,
               (content scan)     Bedrock, Mistral, Groq, Cerebras,
               (audit log)        Together, Fireworks, Ollama, ...)
               (usage tracking)

Three integration modes:

Mode How Best for
Hooks PreToolUse/PostToolUse intercepts Claude Code, Gemini CLI
Django proxy Transparent HTTP proxy with policy enforcement Any agent — simple deployment
Go gateway High-performance sidecar with credential injection Enterprise — scales independently

Why Zentinelle?

  • Open source (MIT) — derisks legal friction for enterprise procurement
  • 24 policy evaluators out of the box covering rate limits, tool permissions, model restrictions, network controls, content scanning, and more
  • Real GRC framework mapping — SOC2, GDPR, HIPAA, EU AI Act, NIST AI RMF
  • Multi-tenant from day one — isolated by tenant_id at every layer
  • Built-in AI assistant — chat with Zentinelle to explain policies, suggest configs, summarize compliance status
  • Encrypted-at-rest LLM credentials — Fernet-encrypted provider keys, never log plaintext
  • Live model discovery — auto-populates from provider /models APIs, stays current as new models ship

Get started

Reference

Project