Self-hosted AI Starter Kit
A Docker Compose template to quickly set up a local AI and low-code development environment.
Evidence: README advises copying .env.example and updating secrets/passwords, showing some awareness of sensitive data; but no least-privilege configuration details, no explicit user confirmation mechanism, limited data flow transparency, no mention of dependency security, no description of external effects, rollback only via upgrade commands, source attribution via GitHub link. Deductions: lack of specific security config and user confirmation process, dependency security not mentioned, external effects not described.
Evidence: README provides multiple installation methods (GPU, CPU, Mac), documentation and commands are consistent; dependencies (n8n, Ollama, Qdrant, PostgreSQL) are well-known, but no version pinning or checksums; failure messages not detailed. Deductions: dependency versions not pinned, failure handling guidance insufficient.
Evidence: Clearly targets developers and enterprises, provides multiple use cases (AI Agents, PDF summarization, etc.); capability boundaries via component list; trigger precision not detailed; environment fit covers GPU, CPU, Mac. Deductions: trigger precision not detailed.
Evidence: README structure clear with sections for installation, usage, upgrade, support; installation notes detailed; naming stable (project name consistent); examples and FAQ links provided; known limitations mentioned (not production-optimized); license Apache-2.0; versioning/changelog not provided; maintenance responsibility not explicit. Deductions: versioning/changelog missing, maintenance responsibility not explicit.
Evidence: Output usability via provided workflows and templates; marginal value in quick local AI environment setup; cost-benefit via open-source and local running. Deductions: no performance or cost data.
Evidence: Claims in README (e.g., component capabilities) verifiable via links; but no independent verification sources; facts and inferences not clearly separated. Deductions: lack of independent verification and fact/inference separation.
- Dependency versions are not pinned, which may introduce incompatibility or security risks.
- No user confirmation mechanism is provided, sensitive operations may execute automatically.
- No versioning or changelog is provided, making updates hard to track.
What does this agent do, and when should you use it?
The Self-hosted AI Starter Kit, curated by n8n, is an open-source Docker Compose template designed to rapidly initialize a comprehensive local AI and low-code development environment. It bundles self-hosted n8n (a low-code platform with over 400 integrations), Ollama (for running local LLMs), Qdrant (a vector database), and PostgreSQL. The kit pre-configures networking and storage, supports Nvidia GPU, AMD GPU (Linux), Apple Silicon (Mac), and CPU profiles, and includes an example workflow. It is intended for quick starts and proof-of-concepts, not production use.
This kit orchestrates multiple containers via Docker Compose: n8n provides a low-code workflow editor with AI nodes; Ollama downloads and runs local language models (e.g., Llama3.2); Qdrant serves as a vector store; PostgreSQL handles data persistence. After installation, users access n8n at http://localhost:5678/, open the included workflow (ID: srOnR8PAY3u4RSwb), and interact with the AI agent via a chat interface. The workflow uses Ollama for local inference and Qdrant for memory. A shared folder is mounted to the n8n container at /data/shared for file access.
- Developers wanting to build AI workflows locally to avoid data privacy issues.
- Business users needing AI agents for scheduling appointments or automating tasks.
- Teams analyzing confidential company PDFs without cloud exposure.
- IT teams enhancing Slack bots for smarter operations.
- Financial analysts processing private documents at low cost.
- Low-code enthusiasts experimenting with self-hosted AI integrations.
What are this agent's strengths and limitations?
- Fast setup: one command launches a complete local AI environment.
- Well-integrated components: n8n, Ollama, Qdrant, PostgreSQL work together seamlessly.
- Over 400 integrations and advanced AI nodes for extensibility.
- Supports multiple hardware profiles (Nvidia, AMD, Mac, CPU).
- Includes example workflow and comprehensive documentation.
- Not optimized for production; intended for proof-of-concept.
- Requires Docker and Docker Compose knowledge.
- First run downloads models, which can be time-consuming.
- Mac users cannot use GPU directly, limiting performance.
- CPU-only mode may be slow for large models.
How do you install or deploy this agent?
1) Clone the repository: git clone https://github.com/n8n-io/self-hosted-ai-starter-kit.git \n2) cd self-hosted-ai-starter-kit \n3) cp .env.example .env and update secrets/passwords \n4) Choose the appropriate command: \n - Nvidia GPU: docker compose --profile gpu-nvidia up \n - AMD GPU (Linux): docker compose --profile gpu-amd up \n - Mac/Apple Silicon: run docker compose up (CPU) or follow docs for GPU \n - CPU (others): docker compose --profile cpu upHow do you use this agent?
1) Open http://localhost:5678/ in your browser to set up n8n (first time only). \n2) Open the included workflow: http://localhost:5678/workflow/srOnR8PAY3u4RSwb \n3) Click the Chat button at the bottom of the canvas to start. \n4) Wait for Ollama to download Llama3.2 (check Docker logs). \n5) Use shared folder path /data/shared for file operations.
How does this agent compare with similar options?
The n8n platform within the kit competes with automation tools like Zapier and Make, but emphasizes self-hosting and local AI capabilities.