Governance scaled to
your regulatory load.
Every tier includes the DREP lifecycle. Volume, enforcement depth, and deployment isolation determine the fit.
Observer
AI First Team · Startup
- Up to 3 governed runtimes
- First 2M events included
- Divergence monitor + drift scoring
- Basic provenance export
- CT-Toolkit integration
Enforcer
BFSI · Mid-Market
- Unlimited governed runtimes
- First 10M events included
- Runtime enforcement + policy review
- 3rd Party Integrations (ServiceNow, etc)
- Adv. Audit Reports (EU AI, SoC, NIST, etc)
Most adopted
Auditor
Regulated BFSI · Compliance-Heavy
- All Enforcer features
- First 25M events included
- BYOK for LLM Models
- SSO / SAML 2.0
- Single-Tenant Cloud
Enterprise & Gov
Enterprise SLA · Global Scale
- Air-Gapped & On-Premise
- Zero-knowledge governance
- Unlimited events included
- TG-Sentinel: Ultra (27B) model access
- Custom compliance & integrations
- Dedicated 24/7 support & SLA
Stateless filters vs.
stateful governance.
| Feature | NeMo Guardrails | Azure AI Safety | Credo AI / LlamaGuard | Theseus Guard |
|---|---|---|---|---|
| Architecture | Stateless (Legacy) | Stateless (Legacy) | Stateless (Legacy) | Stateful (TG-Core) |
| Multi-turn Memory | No | No | No | Yes |
| Runtime Enforcement | No | Partial | No | Yes |
| Agent Identity Tracking | No | No | No | Yes |
| Tamper-Evident Provenance | No | No | No | Yes |
| Agentic / Nested Agents | No | No | No | Yes |
| BFSI / Regulatory Focus | No | No | Partial | Yes |
| Open-Source Foundation | Yes | Yes | No | Yes |