LLMGW application
LLMGW is built as docker containers which can be deployed on Azure or AWS providing a cloud-native, scalable foundation with enterprise-grade security and operational capabilities. The platform consists of four primary container applications that work together to deliver comprehensive AI gateway functionality.
LLMGW proxy
Function: Core API gateway handling all AI provider communications
Features: Request routing, load balancing, authentication, and response processing
Scaling: Available auto-scaling based on request volume and latency requirements
Access: Internal application only allowed to be accessed from selected networks with all requests being authenticated
Admin Portal
Function: Web-based management interface for platform configuration
Features: User management, project configuration, budget controls, and system monitoring
Access: Secure web interface secured by SSO access with RBA
Grafana Dashboards
Function: Advanced analytics and visualization platform
Features: Cost monitoring, alerting, and performance metrics visualization
Integration: Connected to all telemetry and metrics data sources
Access: Secure web interface secured by SSO access supporting RBAC
Open Telemetry Metrics Emitter
Function: Standardized observability and metrics collection
Features: Distributed tracing, custom metrics, and performance monitoring
Standards: Full Open Telemetry compliance