Dify AIO: An All-in-One LLM Application Platform for Home Labs

This article introduces the dify-aio project, which packages the complete Dify platform into a single container. Designed specifically for Unraid users, it includes built-in Postgres/pgvector, Redis, and a sandbox environment, enabling individual developers to quickly deploy AI workflows without Docker Compose—ideal for home lab scenarios.

As an open-source LLM application development platform, Dify offers complete features from prompt orchestration and RAG knowledge bases to Agent workflows. However, the official recommended Docker Compose deployment requires managing multiple service containers (Postgres, Redis, Weaviate, etc.), which has high maintenance costs—especially on single-container NAS systems like Unraid. Thus, dify-aio was born: it packages the entire Dify ecosystem into an All-in-One container, allowing you to launch the full platform in minutes without complex configurations.

The dify-aio single container includes the following core components:

1. Dify main service: Web frontend, API service, task queue, supporting multi-tenancy, permission management, version control, visual dialogue flow design, knowledge base building, etc.
2. Postgres+pgvector: Stores application data, user sessions, and vector embeddings—enables RAG functionality without external vector databases.
3. Redis: Cache layer and message queue, supporting SSE real-time push and asynchronous tasks.
4. Sandbox environment: Isolates code execution for safe running of Python code, API calls, or file processing.
5. Nginx reverse proxy: Unified entry point, handling static resources, API routing, and load balancing.

Deployment Methods

• Unraid deployment: Search for "dify-aio" in the Community Applications plugin, configure environment variables, then start with one click (preset resource limits and storage mappings included).
• General Docker deployment: Pull the image and run (command see original text). Use Cases
• Personal knowledge base: Convert local documents to vector databases, and use private LLMs for data-local Q&A.
• AI workflow automation: Visually orchestrate LLM calls, API requests, and conditional logic.
• Multi-agent collaboration: Create professional agents and coordinate execution via parent agents.
• Rapid prototype validation: Developers quickly validate LLM application ideas.

Environment Variables

Variable Name	Description	Default Value
INIT_PASSWORD	Initial administrator password	random
DB_PASSWORD	Postgres password	difyai123456
REDIS_PASSWORD	Redis password	difyai123456
SANDBOX_API_KEY	Sandbox API key	random
Storage Mapping: It is recommended to map /app/data to a host persistent directory, which includes postgres/, redis/, uploads/, and logs/.
Resource Requirements: Minimum 2 CPU cores + 4GB memory; recommended: 4 CPU cores + 8GB memory (supports concurrent RAG and code execution).

Feature	dify-aio	Official Docker Compose
Deployment Complexity	Single container one-click start	Multi-service orchestration required
Maintenance Cost	Low (automatic internal coordination)	High (monitor status of each service)
Use Cases	Individual/small team	Enterprise/production environment
Scalability	Vertical scaling (hardware upgrade)	Horizontal scaling (multi-instance cluster)
Customization Freedom	Limited by preset configurations	Fully customizable

dify-aio reflects the trend of LLM tools becoming more accessible to the public, encapsulating enterprise-level features into a form accessible to individuals. It does not replace the official architecture but provides a lightweight option for home labs, rapid validation, education, and other scenarios. Future plans may include GPU inference acceleration, built-in model management interface, integration with more open-source tools, and more.