Launched May 20, 2025
Harbor
Effortlessly run LLM backends, APIs, frontends, and services with one command.
FreeAICLIOpen Source2,577 impressions#5 of its week
Harbor is a containerized LLM toolkit that helps you to run LLMs and satellites via a concise CLI
- Convenience utils (manage models, config, debug, URLs, tunnels, more)
- Built-in optimizing proxy and benchmarking
- Service CLIs without install via Docker
- Shared caches
- Config profiles, history
Harbor is a CLI tool that orchestrates containerized LLM backends, frontends and services with a single command.
- for
- Developers who want to run local LLM stacks quickly without manual Docker setup
- pricing
- open source
- license
- Apache-2.0
- v0.5.14v0.5.14 - Latitude, AutoGen Studio, LobeHub, R2R, Caddy, RAGFlow, ClaraVerse, Qw18 days ago
- v0.5.1323 days ago
- v0.5.12v0.5.12 - PrismML llama.cpp25 days ago
works withOllamallama.cppvLLMOpen WebUISearXNGSpeachesComfyUI
Key features
6 features of Harbor
- Single-command stack launch — Run multiple LLM backends, frontends and services together with `harbor up`.
- Containerized LLM toolkit — Provides Docker-based backends like Ollama, llama.cpp, vLLM.
- Convenience utilities — CLI helpers for model management, config, debugging, URLs and tunnels.
- Optimizing proxy & benchmarking — Built-in proxy for performance tuning and tools to benchmark models.
- Service CLIs via Docker — Run service CLIs without installing them locally, using Docker containers.
- Shared caches & profiles — Caches models across services and supports config profiles and history.
Use cases
- Spin up a full local LLM stack with Open WebUI and voice chat in seconds
- Benchmark different LLM backends on the same hardware
- Add web search (SearXNG) or image generation (ComfyUI) to a running model
- Develop and test LLM-driven applications without manual Docker compose files
Harbor vs alternatives
 Harbor |  Daytona |  BuildShip | .png?width=64) Langfuse |  Spawned | |
|---|---|---|---|---|---|
| Best for | One-command local LLM stack orchestration | One-command dev environment provisioning | AI backend visual builder | LLM app observability | Self-hosted cloud deployment |
| Pricing | Open source | Free | Free | Subscription | Free |
| DevHunt upvotes | 1 | 339 | 118 | 89 | 68 |
| Launched | May 2025 | May 2024 | Jan 2023 | Jan 2023 | Jul 2026 |
- Harbor vs Daytona: Daytona focuses on generic dev environment setup, not LLM stack orchestration
- Harbor vs BuildShip: BuildShip provides a visual AI backend builder, whereas Harbor is CLI-driven and container-based
- Harbor vs Langfuse: Langfuse offers observability for LLM apps, not the orchestration of services
- Harbor vs Spawned: Spawned is a Vercel-like cloud platform, while Harbor runs locally via Docker
Harbor FAQ
How do I start a Harbor stack?+
Run `harbor up` followed by the names of the services you want, e.g. `harbor up searxng speaches`.
Do I need Docker installed?+
Yes, Harbor uses Docker Compose to orchestrate the containers for backends and services.
Can I run multiple services together?+
Yes, Harbor wires backends, frontends and auxiliary services together automatically.
Is Harbor free to use?+
Harbor is an open-source project hosted on GitHub and can be used without cost.
Summarized by DevHunt from github.com · Sep 29, 2026. Details may change; check the official site.
