Features
Detailed documentation for each LLM Gateway capability - Routing, Token Saving, Guardian Agent, Guardrails, Prompt Store, Identity Aware, and more.
Request Routing
Multi-provider load balancing and failover behind a single API key.
Token Saving
Reduce token usage by optimizing input content before it reaches the model.
Guardian Agent
Guide model behavior with gateway-side policy checks for dependency safety and task adherence.
Security Guardrails
Detect and act on sensitive data and adversarial inputs.
Custom Intents
Define your own detection categories with examples.
Rate Limits
Control request rates, token budgets, and key expiry.
Prompt Store
Manage and version system prompts centrally, then reference one or more of them and add inline instructions at request time.
Red Team Testing
Test your guardrails configuration against adversarial prompts.
Identity Aware
Authenticate and track users behind each API key.
SDK Mode
Scan content directly from your application code - no LLM proxy required.
Conversation Grouping
Group related requests together so logs and analytics can be viewed at the conversation level.
Copilot Studio
Connect Microsoft Copilot Studio external threat detection to QuilrAI guardrails.
Self-Service
Let your developers get gateway access and manage their own API keys, without an admin minting a key for every person.
Audit Log
A complete, versioned history of every configuration change to an LLM Gateway app - with one-click rollback and an approval queue for self-service change requests.
Agent Monitoring
Correlate every gateway LLM call with the agent run, trace, workflow, or conversation that produced it - using standard distributed-tracing headers, Quilr agent headers, or provider request-body metadata. No Quilr SDK required.