Does this include an AI model or run one for me?

No. There is no GPU and no bundled model, so Open WebUI does not run a model itself. It is the chat interface. You supply your own OpenAI or Anthropic-compatible API key, or point it at your own Ollama server, and your chosen provider bills you for usage.

How do I connect my AI provider?

After your first login, open Settings, then Connections (or Admin Settings), and paste your provider's API base URL and key, or your Ollama server URL. Pick a model and start chatting. You can add several providers and switch between models per chat.

Where are my chats and uploaded documents stored?

Everything stays in a single SQLite database on your instance volume, backed up nightly and served from the region you choose. Nothing is sent to us for training or profiling. Only the model requests you make go to the AI provider whose key you configured.

Who keeps Open WebUI updated?

We do. Curated updates roll out shortly after upstream releases and are tested on our platform first; security fixes move fast. If an update changes behaviour in a breaking way, we email you before it lands.

What happens to my data if I cancel?

Take a final export or SFTP copy any time before you cancel. After termination we keep your data for 14 days (in case you change your mind), then purge it, backups included.

Can I use my own domain?

Yes. Point a CNAME at your instance, prove ownership with a TXT record, and TLS certificates are issued automatically. Cloudflare-proxied domains work; if you use CAA records, allow letsencrypt.org.

Managed Open WebUI hosting

Your own ChatGPT-style workspace, powered by the AI keys you already pay for

Live in minutes
Daily backups, one-click restore
SFTP access, no lock-in
from US$1.70/mo in Poland

Deploy Open WebUIStart free with $5 credit · no card needed

Open WebUIWhat you're getting

Open WebUI is the popular self-hosted, ChatGPT-style interface for large language models. It gives your whole team one clean, familiar chat workspace with conversation history, folders, prompt presets, document chat, image input, web search, and multi-user accounts with roles, all running on infrastructure you control instead of a seat-priced SaaS.

This is a bring-your-own-key product. There is no GPU here and no model is bundled, so Open WebUI does not run an AI model itself. After your first login you open Settings and paste an OpenAI or Anthropic-compatible API base URL and key, or point it at your own Ollama server, then pick a model. Every reply is billed by that provider, on your account, and you can add several providers and switch models per chat.

Replaces ChatGPT Plus: Instead of paying per seat for one vendor's hosted assistant, you run the polished chat front end yourself and connect the AI provider you already pay for, so your prompts and files stay on infrastructure you control.

On CoreHost

Bring your own key: connect OpenAI, any OpenAI-compatible endpoint, or your own Ollama server, no lock-in to one vendor
Familiar ChatGPT-style UI with history, folders, prompt presets, document chat, and image input
Multi-user with roles: the first visitor becomes owner, then you invite teammates and choose who can use which models
Private by design: chats, uploaded documents, and settings live in a single SQLite database on your own instance, in the region you choose

Included with every appStandard, not add-ons

Daily backups, one-click restore

Every instance is backed up daily. Restore any snapshot yourself from the panel. No support ticket, no fee.

SFTP file access

Direct SFTP access to your app's data folders. Your files are never locked away from you.

Custom domain + auto TLS

Bring your own domain with a CNAME and a TXT verification record. Certificates are issued and renewed automatically.

AI + human support

A 24/7 AI engineer that can actually inspect and fix your app, with a human escalation path. App-level help included.

Automatic updates with rollback

Each update is tested on a copy and backed up first, and rolled back automatically if the app fails its checks.

Per-app firewall

Restrict access to the IP addresses you choose, with an Add my IP button. Off by default.

Backups to your own S3

Add encrypted, incremental backups to your own S3-compatible storage on top of the daily ones we include.

Resize any time

Move up or down a tier from the panel. Your balance is drawn by the hour, so you pay the new rate from the change.

RegionsWhere Open WebUI runs

Open WebUI can be deployed in Sydney, the US and Poland. Pick the region closest to the people who use it. Daily backups are encrypted and stored in the EU.

Sydneyfrom US$2.30/moClosest to Australia and New Zealand.
The USfrom US$2.30/moClosest to the Americas.
Polandfrom US$1.70/moLowest price, 25% less than the other regions.

Moving to CoreHostMove your existing Open WebUI here

Already running Open WebUI on PikaPods, a VPS or your own server? You can bring your data with you. Deploy the app here first, then follow the route below, and keep your old app until you are happy.

Tell us what you are moving

Open WebUI starts empty here. To move chat history and settings from an existing Open WebUI, open a ticket: it needs the database, your uploads and the secret key moved together.

Ways to move it

Do it yourself. Deploy Open WebUI, let it start once, then use the steps on this page, with SFTP and the file manager in the Files tab. A backup taken first gives you a clean point to go back to.
Ask us to move it. Open a ticket from your control panel with the app and where it runs now. Support will tell you what can be moved and help you do it, and nobody touches your old app.

Coming from PikaPods? Switch on SFTP in the pod's settings, stop the pod and copy its files; PikaPods documents this as its manual backup for migrations.

How moving to CoreHost works CoreHost compared with PikaPods

DeployPick a size and a region

1. Plan (Open WebUI needs Standard or larger to run well)

Prices exclude GST/VAT where applicable. Drawn hourly from your prepaid balance, cancel any time.

Too small for this app

Lite

$1.50/mo

1 vCPU
1 GB RAM
5 GB NVMe

Lightweight tools: bookmarks, feeds, uptime checks, password vaults.

Standard

$2.30/mo

1 vCPU
2 GB RAM
10 GB NVMe

The right fit for most apps: automation, wikis, databases.

Power

$3.90/mo

2 vCPU
4 GB RAM
20 GB NVMe

Busy teams and heavier workloads with room to breathe.

Scale

$6.30/mo

3 vCPU
8 GB RAM
40 GB NVMe

The most demanding apps in the catalog, at full stride.

Extra storage+25 GB NVMe per block, stackable on any tier

+$0.00/mo

2. Region

…

Australia

$2.30/mo

…

United States