What is Harbor used for?

Harbor is a CLI and companion app that starts a complete local LLM stack with a single command, wiring together backends, chat front ends, and supporting services. It is a good fit for homelab AI setups, and trying out local models and interfaces quickly.

Is Harbor open source?

Yes. Harbor is open source under the Apache-2.0 licence. Its source code is on GitHub at av/harbor and is written mainly in Python.

Yes. Harbor is open source, so the software itself is free to use.

What are some alternatives to Harbor?

Similar open-source tools in the AI Infrastructure category include ODS, llama.cpp and Unsloth. SaaS products in the same category include AI Stats, Amazon Bedrock and AnyRouter.

Is Harbor actively maintained?

Yes. The most recent commit to Harbor was on 25 September 2026, and the latest release is v0.5.14, published on 25 September 2026. The project has 3.2k stars on GitHub.

Harbor: open-source AI infrastructure

About Harbor

Harbor is a command-line tool, with a companion app, that sets up a local large language model stack for you. Instead of configuring each piece separately, you start the services you want from one command and Harbor handles the Docker Compose wiring between them. The project is released under the Apache-2.0 license.

A stack can combine inference backends such as Ollama, llama.cpp, or vLLM with front ends such as Open WebUI, plus supporting services like SearXNG for web search. The project advertises hundreds of services that can be explored this way. Because everything runs through Docker, the setup is aimed at homelabs, local servers, and developers experimenting with self-hosted AI.

The tool is written in Python and distributed through npm and PyPI. It suits people who want to try different models, interfaces, and tools without maintaining their own compose files, and who prefer to keep data and prompts on hardware they control instead of using a hosted AI service.

Key features

One-command startup of a full local LLM stack
Backends including Ollama, llama.cpp, and vLLM
Front ends such as Open WebUI
Supporting services like SearXNG web search
Docker Compose based service wiring
CLI plus companion app

Good fit for

→Homelab AI setups
→Trying out local models and interfaces quickly

Built with: Python; Docker
Tags: llm; local-ai; docker; self-hosted; ollama; cli; homelab; mcp

Harbor: questions and answers

What is Harbor used for?: Harbor is a CLI and companion app that starts a complete local LLM stack with a single command, wiring together backends, chat front ends, and supporting services. It is a good fit for homelab AI setups, and trying out local models and interfaces quickly.
Is Harbor open source?: Yes. Harbor is open source under the Apache-2.0 licence. Its source code is on GitHub at av/harbor and is written mainly in Python.
Is Harbor free?: Yes. Harbor is open source, so the software itself is free to use.
What are some alternatives to Harbor?: Similar open-source tools in the AI Infrastructure category include ODS, llama.cpp and Unsloth. SaaS products in the same category include AI Stats, Amazon Bedrock and AnyRouter.
Is Harbor actively maintained?: Yes. The most recent commit to Harbor was on 25 September 2026, and the latest release is v0.5.14, published on 25 September 2026. The project has 3.2k stars on GitHub.