Deploy and Host Open WebUI with Ollama Cloud on Railway

Open WebUI chat interface

Open WebUI is the most popular self-hosted AI chat interface, with 156,000+ GitHub stars: a ChatGPT-style app with chat history, documents, web search, image generation and multi-user accounts. This template connects it to Ollama Cloud, so you get open models like gpt-oss, DeepSeek, Qwen and Gemma without running a GPU. ChatGPT Business costs $20–25 per seat — host Open WebUI on Railway for ~$5–10/month plus the model plan you choose.

What This Template Deploys

Service	Purpose
open-webui	Open WebUI (`ghcr.io/open-webui/open-webui:main`, port 8080, public HTTPS domain), preconfigured with `OLLAMA_BASE_URL=https://ollama.com`.
Persistent Volume	Mounted at `/app/backend/data` — users, chats, uploaded documents, settings and the SQLite database survive redeploys.

One service and no GPU. Models run on Ollama's cloud; Open WebUI handles accounts, history, RAG over your files and the chat experience.

About Hosting Open WebUI

Self-hosted AI chat usually means one of two hard paths: renting a GPU to run models yourself, or wiring Open WebUI to paid APIs and managing keys. Ollama Cloud sits in between — open models, hosted for you, through the same Ollama API Open WebUI already speaks.

On Railway, Open WebUI runs with persistent storage for every account and conversation, behind HTTPS, and points at Ollama Cloud from the start. You can add OpenAI-compatible providers, OpenRouter or your own Ollama server later from the admin settings.

Sign up first — the first account becomes admin. Create your account right after deploying, then restrict sign-ups. Also add a WEBUI_SECRET_KEY variable: without a persistent key, everyone is logged out each time the container is recreated. Ollama Cloud needs an API key from ollama.com.

Typical cost: ~$5–10/month on Railway, plus Ollama Cloud (free starter credits, or Pro at $20/month with $60 of usage). ChatGPT Business is $20 per seat billed annually ($25 monthly); Claude Team standard seats are $20–25.

Deploy Open WebUI in Under 5 Minutes

Click Deploy on Railway — Open WebUI starts on a volume (~2–3 minutes).
Open the public URL and create your account immediately; it becomes the admin.
Create an API key at ollama.com and add it to the Ollama connection in Admin Settings → Connections.
Add WEBUI_SECRET_KEY (a long random value) to the service and redeploy so logins survive restarts.
Pick a cloud model, start chatting, and invite your team from Admin Settings → Users.

No GPU. No model downloads. No per-seat subscription.

Common Use Cases

Private ChatGPT alternative — "self hosted chatgpt" for a team without per-seat pricing
Open models without a GPU — gpt-oss, DeepSeek, Qwen and Gemma through Ollama Cloud
Chat with documents — upload files and ask questions with built-in RAG
One interface for many providers — add OpenAI, OpenRouter or your own Ollama next to Ollama Cloud
Team AI workspace — user accounts, roles and shared prompts
Pairs with LibreChat or n8n — use Open WebUI as the chat front end for your AI stack

Configuration

Variable	Required	Description
`OLLAMA_BASE_URL`	✅ Set	`https://ollama.com` (Ollama Cloud). Point it at your own Ollama server instead if you have one.
`WEBUI_SECRET_KEY`	Recommended	Add a long random value so sessions survive container restarts.
`ENABLE_SIGNUP`	Optional	`true` by default — set `false` after your team has joined.
`OPENAI_API_BASE_URL`, `OPENAI_API_KEY`	Optional	Add an OpenAI-compatible provider alongside Ollama.

Ollama Cloud needs an API key. Requests to ollama.com are authenticated, so models won't load until you add your key to the connection in the admin settings.

Open WebUI vs. Alternatives

	Open WebUI on Railway	ChatGPT Business	Claude Team	LibreChat (self-hosted)
Monthly cost	~$5–10 + Ollama plan	$20–25 per seat	$20–25 per seat	Free, you host it
Self-hosted	✅ Yes	❌ No	❌ No	✅ Yes
Per-seat pricing	✅ None	❌ Per seat	❌ Per seat	✅ None
Open models without a GPU	✅ Ollama Cloud	❌ OpenAI models	❌ Claude models	⚠️ Via API providers
Chat with documents (RAG)	✅ Built in	✅ Built in	✅ Built in	✅ Built in
Data stays on your infra	✅ Chats and files	❌ No	❌ No	✅ Yes
License	⚠️ Open WebUI License (branding terms)	❌ Proprietary	❌ Proprietary	✅ MIT

Dependencies for Open WebUI Hosting

Railway account — Hobby ($5/month) runs Open WebUI for a small team.
An Ollama account and API key (free tier, Pro $20/month or Max $100/month).
Optional: OpenAI, OpenRouter or another provider key.

Deployment Dependencies

Open WebUI GitHub — source and releases
Open WebUI Docs — configuration and hardening
Ollama Cloud — cloud models and API keys
Railway Volumes Documentation — persistence

Implementation Details

The service runs the official ghcr.io/open-webui/open-webui:main image with all application data — SQLite database, uploads and vector store — under /app/backend/data on the volume. OLLAMA_BASE_URL points at Ollama Cloud, so model calls go to ollama.com with your API key.

For a team setup, add WEBUI_SECRET_KEY, turn off open sign-ups after onboarding, and consider pinning a specific Open WebUI version instead of main so upgrades happen when you choose.

Frequently Asked Questions

What is Open WebUI? A self-hosted, ChatGPT-style web interface for AI models, with accounts, chat history, document chat and support for Ollama and OpenAI-compatible providers.

How much does it cost compared to ChatGPT Business? About $5–10/month on Railway plus your Ollama plan (free starter credits, or $20/month Pro). ChatGPT Business is $20 per seat per month billed annually, so a 10-person team pays $200+.

Why does Open WebUI show no models? Ollama Cloud requires an API key. Create one at ollama.com and add it to the Ollama connection in Admin Settings → Connections, or point OLLAMA_BASE_URL at your own Ollama server.

Do I lose chats when I redeploy? No. Chats, users, documents and settings live on the /app/backend/data volume. Set WEBUI_SECRET_KEY so people stay logged in across redeploys.

Can I use OpenAI or OpenRouter too? Yes. Add OPENAI_API_BASE_URL and OPENAI_API_KEY, or add connections in the admin settings, and pick models from any provider in the same chat.

Who becomes admin? The first account created after deploy. Sign up immediately, then turn off open sign-ups so only people you invite can join.

Why Deploy and Host Open WebUI with Ollama Cloud on Railway?

Railway is a singular platform to deploy your infrastructure stack. Railway will host your infrastructure so you don't have to deal with configuration, while allowing you to vertically and horizontally scale it.

By deploying Open WebUI with Ollama Cloud on Railway, you get a private, multi-user ChatGPT-style app running open models with no GPU, for a small hosting cost plus the model plan you pick.