LiteMaaS - Models as a Service Platform

LiteMaaS is a proof-of-concept application for managing LLM subscriptions, API keys, and usage tracking. It seamlessly integrates with LiteLLM to provide a unified interface for accessing multiple LLMs with comprehensive budget management.

🎟️ Key Features

• Self-Service: Users can independently manage subscriptions and API keys
• Multi-Model Access: One API key can access multiple models
• Budget Control: Set spending limits at user, team, and API key levels
• Usage Analytics: Comprehensive tracking with dual views:
  • User View: Personal usage, token consumption, and costs
  • Admin View: System-wide analytics with trends, multi-dimensional filtering (users/models/providers/API keys), and data export
• Role-Based Access Control: Three-tier hierarchy (admin/adminReadonly/user) with OpenShift group integration
• Restricted Access Models: Models can be restricted to admin approval.
• Enterprise Security: OAuth2 authentication with JWT tokens and audit logging
• Easy Deployment and Usage: Automated deployment on OpenShift with PatternFly 6 UI

📸 Screenshots

Note: Screenshots below show core functionality. Admin-specific features (,model, subscription, user management, tools,...) are available but not captured in screenshots.

Model Discovery

Subscriptions Management

API Keys Management

API Key Details

Chatbot Playground

Usage Analytics (Admin View shown)

And more...

🚀 Deployment

LiteMaaS supports multiple deployment methods:

OpenShift/Kubernetes Deployment (Recommended for Production)

Quick OpenShift Deployment: Readme

Deploy to OpenShift or Kubernetes using the provided manifests and Kustomize configuration:

# Deploy to OpenShift
oc apply -k deployment/openshift/

# Access at:
# - LiteMaaS: https://litemaas-<namespace>.<cluster-domain>
# - LiteLLM UI: https://litellm-<namespace>.<cluster-domain>

📚 Complete Guide: OpenShift Deployment Guide

Container Deployment (Development/Testing)

Deploy using containers with Docker Compose for local development:

# Local deployment with containers
docker-compose up -d

# Or with Podman
podman-compose up -d

For detailed container deployment instructions, see Container Deployment Guide.

Manual Deployment

For traditional deployment instructions, see the Production Guide.

Quick Production Build

# Build both packages
npm run build

# Build output locations:
# - Backend: backend/dist/
# - Frontend: frontend/dist/

Container Image Build

Build container images for deployment using the automated build script:

📦 Custom Registry: To use a different container registry, edit the REGISTRY variable in scripts/build-containers.sh before building:

# Change this line in scripts/build-containers.sh:
REGISTRY="your-registry.com/your-org"

# Build both backend and frontend container images
npm run build:containers

# Build and push to registry (default: quay.io/rh-aiservices-bu)
npm run build:containers:push

# Push existing images to registry
npm run push:containers

Features:

• 🏷️ Centralized versioning - Uses version from root package.json
• 🐳 Multi-runtime support - Works with Docker or Podman
• 🚀 Flexible workflow - Separate build and push operations
• 🏗️ Platform support - Build for different architectures
• 🔧 Registry flexibility - Easy to configure for any container registry

Script options:

# Build without cache
./scripts/build-containers.sh --no-cache

# Build for ARM64
./scripts/build-containers.sh --platform linux/arm64

# Build with local tags only (no registry prefix)
./scripts/build-containers.sh --local

📚 Documentation

Core Documentation

• 🏗️ Architecture Overview - System design and data flows
• 📖 API Reference - Complete REST API documentation
• 🗂️ Project Structure - Complete file organization map
• 🔧 Configuration Guide - Environment variables and setup
• 🛠️ Development Guide - Local development setup
• 🚀 Production Deployment - Production deployment options
• 🎨 UI Guidelines (PatternFly 6) - Frontend development standards

Feature Documentation

• 📊 Admin Usage Analytics - Comprehensive system-wide analytics with caching
• 🔐 RBAC & User Roles - Three-tier role hierarchy and permissions
• 🔑 Multi-Model API Keys - API key architecture

Development

🚀 Development Quick Start

# Install dependencies
npm install

# Start development servers (both backend and frontend)
npm run dev

# Backend only (http://localhost:8081)
npm run dev:backend

# Frontend only (http://localhost:3000)
npm run dev:frontend

📋 Prerequisites

• Node.js 18.x or 20.x
• PostgreSQL 12+
• npm 8+
• LiteLLM instance (optional, has mock fallback)

🏗️ Architecture

LiteMaaS is a monorepo using npm workspaces with two main packages:

• Backend (@litemaas/backend): High-performance Fastify API server with PostgreSQL
• Frontend (@litemaas/frontend): Modern React application with PatternFly 6 UI

🎯 System Architecture

graph TB
    subgraph "Frontend"
        UI[React + PatternFly 6]
    end

    subgraph "Backend Services"
        API[Fastify API Server]
        Auth[OAuth2/JWT Auth]
        DB[(PostgreSQL)]
    end

    subgraph "External Services"
        LiteLLM[LiteLLM Gateway]
        OAuth[OpenShift OAuth]
    end

    UI <--> API
    API <--> Auth
    API <--> DB
    API <--> LiteLLM
    Auth <--> OAuth

    style UI fill:#e3f2fd
    style API fill:#fff3e0
    style LiteLLM fill:#f3e5f5

Loading

🛠️ Development Setup

1. Clone the repository

   git clone https://github.com/rh-aiservices-bu/litemaas.git
   cd litemaas

2. Set up environment variables

   # Copy example environment files
   cp backend/.env.example backend/.env
   cp frontend/.env.example frontend/.env

3. Configure your environment

   • See Configuration Guide for all available options
   • Minimum required: DATABASE_URL, JWT_SECRET, OAuth credentials

4. Start PostgreSQL (using Docker)

   docker compose -f dev-tools/compose.yaml up -d postgres

5. Run the application

   npm run dev

🧪 Testing

First time running backend tests? You need to set up the test database first:

# Create test database
psql -U pgadmin -h localhost -p 5432 -d postgres -c "CREATE DATABASE litemaas_test;"

# Initialize schema and seed data
cd backend && npm run test:db:setup

Run tests:

# Run all tests
npm run test

# Backend tests
npm run test:backend

# Frontend tests
npm run test:frontend

# E2E tests
npm run test:e2e

# Performance tests
npm run test:perf

Backend integration tests use a separate litemaas_test database to prevent contamination of development data. See Development Setup Guide for details.

🤝 Contributing

We welcome contributions! Please see our Contributing Guide for details on:

• Code style and conventions
• Development workflow
• Submitting pull requests
• Reporting issues

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🔗 Related Projects

• LiteLLM - The AI model proxy that LiteMaaS integrates with
• PatternFly - The UI framework used in the frontend

💬 Getting Help

• 📖 Documentation - Comprehensive guides and references
• 🐛 Issue Tracker - Report bugs or request features
• 🤝 Contributing - Join the development community

---

Built with ❤️ by the CAI Team