SUSE AI Universal Proxy

A comprehensive, modular MCP (Model Context Protocol) proxy system that enables secure, scalable, and extensible AI model integrations.

🚀 Key Capabilities

🔄 MCP Proxy Service - Full-featured HTTP proxy for MCP servers with advanced session management, authentication, and protocol translation.

🔍 Network Discovery - Automated network scanning to discover MCP servers, detect authentication types, and assess security vulnerabilities.

📚 Server Registry - Curated registry of MCP Servers, including GitHub, SUSE MCP's, Atlassian, Gitea, and 20+ other popular services (yes you may contribute to the list!).

🔌 Plugin Management - Dynamic plugin system for extending functionality with service registration, health monitoring, and capability routing.

📖 Documentation

QUICKSTART - Get started quickly with SUSE AI Universal Proxy
REGISTRY - Learn about MCP server registry management
SECURITY - Security guidelines and best practices
AUTHENTICATION - Authentication and authorization options

🏗️ Architecture

The system uses a main container + sidecar architecture where services run as coordinated containers within a single Kubernetes pod:

┌─────────────────────────────────────────────────────────────┐
│                    SUSE AI Universal Proxy                  │
│                                                             │
│  ┌─────────────────────────────────────────────────────────┐ │
│  │                   UNIFIED SERVICE                       │ │
│  │                                                         │ │
│  │  • MCP Proxy with session management                   │ │
│  │  • Server registry and discovery                       │ │
│  │  • Plugin management and orchestration                 │ │
│  │  • Authentication and authorization                    │ │
│  │                                                         │ │
│  │              HTTP: 8911 | HTTPS: 3911                  │ │
│  └─────────────────────────────────────────────────────────┘ │
│                                                             │
│  ┌─────────────┐                                            │
│  │   PLUGINS   │                                            │
│  │  (External) │                                            │
│  │             │                                            │
│  │  Variable   │                                            │
│  │   Ports     │                                            │
│  └─────────────┘                                            │
└─────────────────────────────────────────────────────────────┘

🏃‍♂️ Quick Start

Setup SUSE AI Universal Proxy (Helm installation)

Open a local terminal
Clone the repository: https://github.com/suse/suse-ai-up (branch: main)
enter in the folder suse-ai-up
In values.yaml, set:
- service.type: LoadBalancer
- auth.method: development (for no auth)

Install using the helm chart:

helm install suse-ai-up ./charts/suse-ai-up

Wait for the installation to be completed

Get Service IP

kubectl get svc suse-ai-up-service -n suse-ai-up -o jsonpath='{.status.loadBalancer.ingress[0].ip}'

(Alternative) Setup SUSE AI Universal Proxy (in Rancher)

In Rancher, add repository: https://github.com/suse/suse-ai-up (branch: main)
Go to Apps → Charts
Find and install "SUSE AI Universal Proxy"
Click Install and wait for completion

Setup SUSE AI Universal Proxy UI

In Rancher, add repository: https://github.com/suse/suse-ai-up-ext(branch: v0.1.0)
Go to Extensions
Find and install "SUSE AI Universal Proxy"

Verify Installation and Access Swagger Docs

Access API documentation:

http://{IP ADDRESS}:8911/docs/index.html

Local Development

git clone https://github.com/suse/suse-ai-up.git
cd suse-ai-up
go run ./cmd/uniproxy

Universal Proxy require Kubernets so the ideal development way is to deploy the helm chart in kubernetes

📋 Service Overview