Open WebUI CLI Cheatsheet - open-webui Self-hosted AI Platform Full Reference

Open WebUI is a feature-rich self-hosted AI platform that connects Ollama and any OpenAI-compatible API, with built-in RAG, multi-model chat, and an admin console. open-webui is its CLI: serve to start, set port and data dir, or deploy via Docker and Helm. This sheet covers install, launch, and deploy so you can self-host a private AI portal with an admin console.

AI CLI·20 commands·Last updated 2026-09-10

Install and Start 6

pip install open-webui

Install via pip

open-webui serve

Start the server (default 8080)

open-webui serve --port 8088

Start on a custom port

open-webui serve --host 0.0.0.0

Bind externally

DATA_DIR=./data open-webui serve

Start with a data dir

python -m open_webui serve

Start as a module

Update and Maintenance 6

pip install -U open-webui

Upgrade to latest

open-webui serve --help

Start flags

rm -rf ~/.open-webui

Wipe data (use with care)

pip uninstall open-webui

Uninstall the package

UVICORN_WORKERS=1 open-webui serve

Single worker for migrations (upgrade)

open-webui --help

Top-level commands

Deploy and Config 8

docker run -d -p 3000:8080 ghcr.io/open-webui/open-webui:main

Run via Docker

WEBUI_AUTH=False open-webui serve

Single-user mode (no login)

OLLAMA_BASE_URL=https://example.com open-webui serve

Connect remote Ollama

helm repo add open-webui https://open-webui.github.io/helm-charts

Add the Helm repo

kubectl get pods

Check K8s deploy status

DATA_DIR=~/.open-webui uvx open-webui@latest serve

Start via uvx with a fixed data dir

WEBUI_SECRET_KEY=xxx open-webui serve

Set a persistent key to avoid logout

open-webui serve --port 8080

Explicitly set the default port

Tips

Always set DATA_DIR; otherwise uvx temp dirs drop chats and settings when the process ends.
With multiple workers or K8s replicas, scale to one replica for migrations before scaling up, to avoid DB corruption.
Default port 8080; Docker often maps the UI to 3000 with -p 3000:8080.

FAQ

How is Open WebUI different from LM Studio?

Open WebUI is a platform with admin console, RAG, and multi-model chat (server/team oriented); LM Studio is a local desktop inference client. They pair well: Open WebUI connects to LM Studio Ollama endpoint.

Must it connect to Ollama?

No. It can use any OpenAI-compatible API (cloud, vLLM, LM Studio) and has a built-in engine for local RAG.

Official References

Each command links to its official documentation below, so you can verify the latest usage and read deeper.

Open WebUI Docs

Maintained by LaoHand

Publicly updated on Sep 10, 2026, continuously proofread against official docs.

Contact Us

Wrong command or description? Send us corrections, business inquiries or product feedback by email.