self-hosted-ai-starter-kit: a Docker Compose template for quickly initializing a self-hosted local AI and low-code development environment

What it solves

This project provides a pre-configured environment for users to quickly deploy and run AI workflows locally. It eliminates the complex setup process of installing and connecting multiple AI tools (like LLMs, vector stores, and automation platforms) by providing a single Docker Compose template.

How it works

It uses Docker Compose to orchestrate a suite of compatible AI components. The kit includes:

• n8n: A low-code platform used to design and automate AI workflows with over 400 integrations.
• Ollama: A platform for running local LLMs (such as Llama 3.2).
• Qdrant: A high-performance vector store for managing and retrieving AI data.
• PostgreSQL: A database for handling large amounts of data safely.

Depending on the hardware (Nvidia GPU, AMD GPU, or CPU), users can launch the entire stack using specific Docker profiles.

Who it’s for

Developers and AI enthusiasts who want to build and test AI agents, private document analysis tools, or smarter bots without relying on on-cloud AI services, ensuring data privacy and minimal cost.

Highlights

• One-click deployment: Uses Docker Compose profiles for different hardware configurations (Nvidia, AMD, and CPU).
• Local-first approach: Combines local LLMs via Ollama and a vector store via Qdrant to keep data private.
• Low-code automation: Leverages n8n's AI nodes (like AI Agent, Text Classifier, and Information Extractor) to build complex workflows.
• Local file access: Includes a shared folder mount for n8n to interact with the local filesystem.