Harbor
CLI tool that spins up a complete local AI stack with a single command, wiring together 90+ services including LLM backends, frontends, and RAG tools.
About
Harbor is an open-source command-line tool that takes the complexity out of running a local AI environment. A single `harbor up` command deploys pre-configured Docker services covering LLM inference engines such as Ollama, llama.cpp, and vLLM, plus frontends, vector stores, and supporting tools, all connected and ready to use without manual wiring.
It supports over 20 inference backends and includes built-in web search via SearXNG, voice chat via text-to-speech, agentic workflow modules, and direct integration with coding CLIs like Claude Code and GitHub Copilot. Profiles let you save different stack configurations, and an eject feature exports standalone Docker Compose files for portability. Harbor is free under the Apache 2.0 license.
It supports over 20 inference backends and includes built-in web search via SearXNG, voice chat via text-to-speech, agentic workflow modules, and direct integration with coding CLIs like Claude Code and GitHub Copilot. Profiles let you save different stack configurations, and an eject feature exports standalone Docker Compose files for portability. Harbor is free under the Apache 2.0 license.