AI & Machine Learning

The AI package provides a self-hosted AI platform with a chat interface and unified model gateway. It supports both local models (via Ollama) and cloud providers (OpenAI, Anthropic, Google).

Services

Service	Description	Deploy
Open WebUI	ChatGPT-like interface for AI models	./uis deploy openwebui
LiteLLM	Unified API gateway for multiple LLM providers	./uis deploy litellm

Quick Start

./uis stack install ai-local

Or deploy individually:

./uis deploy postgresql   # Required dependency
./uis deploy litellm
./uis deploy openwebui

How It Works

Users → Open WebUI → LiteLLM → Ollama (local models)
                              → OpenAI API
                              → Anthropic API
                              → Google Gemini API

1. Users interact with Open WebUI's chat interface
2. Open WebUI sends requests to LiteLLM's OpenAI-compatible API
3. LiteLLM routes to the appropriate model provider
4. Conversations are stored in PostgreSQL

All services deploy to the ai namespace.

Guides

• Model access configuration — control which models users can see
• LiteLLM client keys — generate API keys for external tools
• Environment management — manage the AI infrastructure