AI Platforms & Chat UIs
Comprehensive AI platforms for building LLM applications, managing multiple models, and providing ChatGPT-like interfaces.Available Services
AnythingLLM
Port: 3115 | Memory: 512 MB | Maturity: StableAll-in-one Desktop & Docker AI application with built-in RAG, AI agents, document chat, and multi-user support.Features:
- Built-in RAG pipeline
- AI agents support
- Document chat
- Multi-user workspace
- Model agnostic
Dify
Port: 3110 | Memory: 1024 MB | Maturity: StableOpen-source LLM app development platform with visual AI workflow builder, RAG pipeline, agent capabilities, and model management.Features:
- Visual workflow builder
- RAG pipeline
- Agent capabilities
- Model management
- API and SDK
Flowise
Port: 3120 | Memory: 256 MB | Maturity: StableDrag & drop UI to build customized LLM flows, chatbots, and AI agents visually.Features:
- Drag-and-drop flow builder
- Pre-built components
- Custom integrations
- LangChain support
- API endpoints
LibreChat
Port: 3090 | Memory: 512 MB | Maturity: StableEnhanced ChatGPT clone supporting multiple AI providers (Claude, GPT, Gemini, local models) with agents and code interpreter.Features:
- Multi-provider support
- ChatGPT-like interface
- Agents and plugins
- Code interpreter
- Conversation management
LiteLLM Proxy
Port: 4000 | Memory: 256 MB | Maturity: StableUnified gateway for 100+ LLM providers with load balancing, fallbacks, spend tracking, and caching.Features:
- 100+ provider support
- Load balancing
- Automatic fallbacks
- Spend tracking
- Response caching
- OpenAI-compatible API
LITELLM_MASTER_KEYDocumentationOpen WebUI
Port: 3080 | Memory: 256 MB | Maturity: StableBeautiful ChatGPT-like web interface for Ollama and other LLM providers. Features RAG, web search, and multi-user support.Features:
- ChatGPT-like UI
- Ollama integration
- RAG support
- Web search
- Multi-user
- Model management
Usage Examples
Add AI platforms to your stack
Use the AI Playground preset
Combine multiple platforms
Platform Comparison
| Platform | Primary Use Case | RAG | Agents | Visual Builder | Memory |
|---|---|---|---|---|---|
| AnythingLLM | Document chat & RAG | ✅ | ✅ | ❌ | 512 MB |
| Dify | Full platform | ✅ | ✅ | ✅ | 1024 MB |
| Flowise | Visual flow building | ✅ | ✅ | ✅ | 256 MB |
| LibreChat | Chat interface | ❌ | ✅ | ❌ | 512 MB |
| LiteLLM | API gateway | ❌ | ❌ | ❌ | 256 MB |
| Open WebUI | Ollama interface | ✅ | ❌ | ❌ | 256 MB |
Architecture Patterns
Full RAG Stack
Multi-Model Gateway
Visual Workflow Platform
Integration Tips
- LiteLLM as Gateway: Use LiteLLM to unify access to multiple providers
- Ollama for Local Models: Pair platforms with Ollama for local inference
- RAG Requirements: Most platforms work best with vector databases (Qdrant, ChromaDB)
- Caching: Add Redis for improved performance
- Multi-User: Configure authentication and user management for production