Open WebUI logo

Deploy Open WebUI

A private AI chat for your team, with your own documents

AIUtilitiesOpen WebUI + PostgreSQL

Deploy Open WebUI with PostgreSQL and pgvector, a persistent volume and a public HTTPS URL, inside its own Kata micro VM. Chat with models from an Ollama app in the same project or any OpenAI-compatible API, and ask questions about the files you upload. The admin account is created at deploy and sign up is closed.

What you get with Open WebUI

  • Your models, one interfaceConnect an Ollama app in the same project at http://ollama, or any OpenAI-compatible API with your own key, and pick a model per chat.
  • Chat with your documentsUpload files to a chat or a knowledge base. They are embedded and stored with pgvector in PostgreSQL, the matching passages are added to every question, and each answer lists its sources.
  • Closed by defaultThe admin account is created on first start from a generated password, so nobody can claim the instance by opening it first. Sign up is off; the admin adds people.
  • Locked downTools, Functions and runtime pip installs, the features that run Python on the server, are switched off by environment variables the admin panel cannot override.
  • Its own micro VMOpen WebUI runs inside a Kata micro VM with its own kernel, not in a container that shares the node’s kernel.
  • History that persistsUsers, chats, settings and document vectors live in PostgreSQL. Uploaded files sit on a volume, so a redeploy keeps everything.

After you deploy Open WebUI

  1. 1
    Sign in as admin

    Open the public URL and sign in with admin@example.com and the password in WEBUI_ADMIN_PASSWORD, from the app’s Variables in the dashboard.

  2. 2
    Connect a model

    Deploy the Ollama template in the same project and its models appear on their own, or add an OpenAI-compatible connection in Admin Panel, Settings, Connections.

  3. 3
    Add your team

    Sign up is closed. Add people in Admin Panel, Users, and give them their first password.

How Open WebUI is wired

Open WebUI and PostgreSQL ship together. The public hostname is the chat app. Models come from an Ollama app in the same project or from an API you choose.

PostgreSQL holds everything
Users, chats, settings and the pgvector embeddings of your documents live in the attached PostgreSQL. Restart the app and everything is still there.
Files sit on a volume
Uploaded files live on a 5 GB volume mounted at /app/backend/data.
Ollama stays private
An Ollama app in the same project answers at http://ollama with no public URL. Open WebUI is the only way in, behind its own login.
Keep the secret key
WEBUI_SECRET_KEY signs every session and is generated at deploy. Changing it signs everyone out.

Why run Open WebUI on StackBlaze

  • Scale your appChange CPU, RAM, and replicas without a rebuild, or let autoscaling add replicas when CPU climbs.
  • Git-based CI/CDPush to GitHub and it builds and ships. Pull requests get their own preview URL on Pro and Business.
  • One-click rollbackA deploy went wrong? Go back to any previous successful deploy.
  • Environments built inProduction, staging, and dev as separate environments in one project, each with its own URL.
  • Built-in observabilityCPU, memory, and response time per app, live logs with search, and alerts to Slack, Discord, or a webhook.
  • Developer-first toolsA CLI and Stack MCP: deploy, read logs, and wire services from your terminal or your AI assistant.
  • Built for teamsInvite teammates with roles. An audit log records who changed what.
  • Managed Postgres with HAPostgres is run for you. Turn on high availability for a three-node cluster when the app needs it.

Common questions about Open WebUI

How do I sign in after deploy?

Open the public URL and sign in as admin@example.com with the password generated in WEBUI_ADMIN_PASSWORD, under the app’s Variables in the dashboard. Change the password afterwards in Settings, Account.

Does it include a model?

No. Deploy the Ollama template in the same project and Open WebUI finds it at http://ollama, or add an OpenAI-compatible API key in Admin Panel, Settings, Connections.

How do document chats work?

Documents are embedded by an embedding model, by default nomic-embed-text on your Ollama app (add it to MODELS_TO_PULL there). You can pick an OpenAI-compatible embedding API instead in Admin Panel, Settings, Documents. Vectors are stored with pgvector in PostgreSQL.

How do I add teammates?

Public sign up is off. The admin adds people in Admin Panel, Users and sets their first password. There is no email, so passwords are reset there too.

Can I use Tools and Functions?

Not in this template. Tools and Functions run Python inside the app, so they are switched off with ENABLE_TOOLS and ENABLE_FUNCTIONS, which the admin panel cannot override. Connections to external MCP and OpenAPI tool servers stay enabled.

Can I use my own domain?

Yes. Attach a custom domain from the dashboard, then set WEBUI_URL to it and add it to CORS_ALLOW_ORIGIN, separated by a semicolon. Custom domains

What does it cost?

You pay for the compute, storage and bandwidth Open WebUI and PostgreSQL use, not per seat or per message. Model usage on a paid API is billed by your provider. Pricing

More AI templates