Policy enforcement
Define and enforce governance policies across every API call. Role-based access, content filtering, and usage constraints applied consistently at the gateway layer.
Rai Shield
Route, govern, and protect when using agents, data, and models. Shield provides a governance-first approach to AI infrastructure, with policy enforcement, audit trails, and data sovereignty built into every request.
Performance that matters
Governance should never be a bottleneck. Shield delivers enterprise-grade policy enforcement at wire speed.
Key features
Routing intelligence
Intelligent model and provider routing based on cost, latency, capability, and policy. Automatic failover and load balancing across providers without vendor lock-in.
Audit trails
Complete observability and audit logging for every request. Structured logs, compliance reporting, and forensic analysis capabilities for regulatory requirements.
Data sovereignty
Control where data flows and ensure regulatory compliance. Geographic routing, data residency enforcement, and jurisdiction-aware processing for sensitive workloads.
Multi-provider support
Connect to any AI provider through a unified interface. Provider-agnostic abstractions ensure portability while governance policies apply uniformly across all providers.
Cost optimisation
Intelligent routing based on real-time pricing, model capabilities, and performance metrics. Reduce AI infrastructure costs by 50-90% through optimal provider selection governed by policy.
What AI governance means
- Routing
- Intelligent model and provider routing based on cost, latency, and capability.
- Policy enforcement
- Define and enforce governance policies across every API call.
- Audit trails
- Complete observability and audit logging for compliance.
- Data sovereignty
- Control where data flows and ensure regulatory compliance.
Deploy in minutes
Single command deployment to any edge platform. No complex configuration, no vendor lock-in.
Edge networks
Major edge platforms including Cloudflare, Fastly, and Vercel. Sub-millisecond startup times with minimal memory usage.
Kubernetes
Enterprise container orchestration with automatic scaling, service integration, and zero-downtime updates.
Bare metal
Direct hardware deployment with intelligent resource management and hardware optimisation for maximum performance.
Serverless
AWS Lambda, Google Cloud Functions, Azure Functions. Optimised for fast cold starts and efficient resource utilisation.
Single binary, zero dependencies, deployed globally in under 60 seconds
Use cases
Regulated industries
Enforce AI governance policies required by financial services, healthcare, and public sector regulations. Complete audit trails and data sovereignty controls.
Enterprise DevOps
Self-hosted deployment with enterprise security, compliance features, and unified governance across multiple teams, projects, and providers.
Edge computing
Ultra-low latency AI processing with governance at the edge. Policy enforcement and audit logging without sacrificing performance.
Research organisations
Complete transparency and control over AI workloads. Audit every request, maintain data sovereignty, and ensure reproducible research outcomes.
Open source
Shield is open source under LGPL and MPL licenses.
Get started with Rai Shield
Join our early access programme to deploy governed, high-performance AI infrastructure. Priority access for organisations committed to responsible AI operations.
Early access includes preview builds, direct engineering support, and influence on feature development priorities.