Harbor: One Command to Deploy Your Complete AI Stack
Harbor lets you deploy a complete pre-wired LLM stack with hundreds of AI services using a single command. Learn about its architecture, features, and available services.
What is Harbor?
Harbor is a CLI and companion application that lets you spin up a complete local LLM stack with a single command. It brings together backends like Ollama, llama.cpp, and vLLM, frontends like Open WebUI, plus supporting services — SearXNG for web search, Speaches for voice chat, ComfyUI for image generation — all pre-wired to work together. No manual setup, no endless configuration files.
🚀 Want to deploy Harbor yourself?
Docker configs, system requirements, and installation guides — all on one page.
View Harbor Tool Page →Why Harbor?
Setting up a local AI stack traditionally means juggling Docker Compose files, environment variables, network configurations, and version mismatches. Harbor eliminates all of that. With over 150 pre-configured services, it's the easiest way to go from zero to a fully functional AI environment.
Key benefits:
- One-command setup —
harbor upstarts everything - Hundreds of pre-wired services — LLMs, vector DBs, MCP servers, agent frameworks, and more
- Automatic cross-service connectivity — Open WebUI talks to Ollama, SearXNG, and Speaches out of the box
- Cross-platform — Works on Linux, macOS, and via the companion desktop app
- Open source — Apache-2.0 licensed with 3.1K+ GitHub stars
Architecture Overview
Harbor operates as an orchestration layer on top of Docker. It generates Docker Compose configurations dynamically based on the services you request, ensuring every service is properly connected to every other service it needs.
Available Services
| Category | Examples |
|---|---|
| LLM Backends | Ollama, llama.cpp, vLLM, DMR, MLX, oMLX |
| Frontends | Open WebUI, ChatNio, ChatUI, SillyTavern |
| Vector Databases | Chroma, Qdrant, Weaviate, Milvus |
| MCP Servers | MCP Inspector, MCP Forge, MetaMCP, MCPO |
| Agent Frameworks | OpenHands, AutoGPT, AgentZero, Codex |
| Image/Video | ComfyUI, Stable Diffusion WebUI |
Key Features in Detail
Harbor Boost
Boost is Harbor's agentic module system. It adds web search, URL reading, and custom tool support to any LLM backend. Features include quick-hop search, deep research workflows, and automated diff analysis — all accessible via harbor launch --workflow.
Harbor Launch
Use Harbor backends from coding and agent CLIs without hand-editing provider configs. harbor launch starts an OpenAI-compatible backend and wires the selected model into tools like Codex, Claude Code, and OpenCode.
🚀 Ready to try Harbor?
Full system requirements, Docker setup, and deployment guides on the tool page.
View Harbor Tool Page →License and Community
Harbor is released under the Apache-2.0 license with over 3,100 stars on GitHub. The community is active on Discord, and the project maintains extensive documentation including installation guides, a full CLI reference, and practical tutorials for local LLM setup, web RAG, coding agents, and more.