Deploy Open WebUI in one click.
The self-hosted AI chat interface. We review it, secure it and run it in production for you. Pick a region in Switzerland, France or Germany, or bring your own machines through a secure tunnel. We keep it patched, backed up and online. You just chat.
★ 136k stars · BSD-3 (variant) · Python · tested versions · shipped with a runbook
What is Open WebUI?
The most popular open-source chat interface for large language models, and one of the most-deployed apps in our catalog.
Open WebUI is an open-source, self-hosted web interface for working with large language models. It connects to Ollama for models running on your own hardware and to any OpenAI-compatible API, so local and hosted models sit side by side in one chat interface. It ships retrieval-augmented generation for chatting with your own documents, multi-user support with role-based access control, and can run entirely offline.
That last part is the point. AI chat traffic is some of the most sensitive data a team produces: prompts, internal documents, embeddings, whole reasoning trails. Self-hosting Open WebUI keeps all of it on infrastructure you control instead of a third-party AI provider’s logs.
The catch: self-hosting means someone has to run it. Reverse proxy, TLS, authentication, upgrades that ship frequently, security patches, backups of every conversation and document your team has fed it. That someone is usually you, at the worst possible time.
Open WebUI, without the ops.
Everything below is what you’d otherwise build and babysit yourself.
Your prompts stay where you put them.
Deploy to ch-zh1 (Zurich), fr-pa1 (Paris) or de-fr1 (Frankfurt). Chat history, uploaded documents and embeddings stay in the region you choose, under Swiss or EU law and in line with GDPR.
Reviewed before it ships.
Our Open WebUI package is reviewed, secured and locked to versions we have tested, and it ships with a runbook. We track CVEs upstream and patch them fast.
Bring your own models.
Run Open WebUI next to Ollama on your own GPU boxes. We drop in a secure tunnel and manage the deployment in place. Your models and your weights never leave your own network.
TLS, upgrades, scaling: handled.
Certificates issued and renewed, rolling zero-downtime updates, instant rollback, health checks. Open WebUI ships new releases fast; we keep you current without breaking your team’s chats.
Backups you can actually restore.
Offsite backups of your conversations, documents and settings. We run restore drills, not just snapshots.
Leave whenever you want.
Open WebUI is open source and so is your exit: your data leaves with you, in usable form. Pilae runs it for you. It never owns it.
Self-hosting Open WebUI: DIY vs Pilae.
Both paths end with Open WebUI in production. Only one of them ends there this afternoon.
| Self-hosted DIY | On Pilae | |
|---|---|---|
| Initial setup | Docker, reverse proxy, DNS, auth. An afternoon if nothing fights back | One click, live in minutes |
| TLS certificates | Issue, wire up, remember to renew | Issued and renewed automatically |
| Upgrades | Frequent releases to track, test and apply yourself | Tested versions, rolled out with instant rollback |
| Security patching | You watch the CVE feed | Upstream tracked, CVEs prioritized |
| Backups | cron + object storage + hope | Offsite, verified with restore drills |
| Data residency | Wherever your VPS happens to be | You pick: Zurich, Paris, Frankfurt, or your own racks |
| 3 a.m. failures | Yours | Ours |
| Multi-user & access | Configure and secure RBAC yourself | Secure defaults; SSO via Keycloak from the same catalog |
| GPU access for local models | Expose your GPU box to the internet, carefully | Secure tunnel; models stay inside your own network |
From catalog to production in three steps.
No servers to set up, no configuration to write. The platform work is already done.
Pick Open WebUI.
Find it in the catalog: reviewed, secured, locked to tested versions, runbook included.
Choose where it runs.
A Pilae region in Switzerland, France or Germany, or your own machine connected through a secure tunnel. Same console either way.
Deploy and connect your models.
One click to production with TLS and backups from day one. Point Open WebUI at Ollama or any OpenAI-compatible endpoint, invite your team, done.
AI chat data is sensitive. Keep it close.
Every Open WebUI deployment is pinned to the region you choose. Encryption, dedicated VLANs, and the option to hold your own encryption keys for regulated workloads. Or skip our regions entirely and run it on-prem.
Deploying Open WebUI: the questions we get.
If yours isn’t here, reach out and we’ll answer it, usually the same day.
What’s the easiest way to self-host Open WebUI?
Deploying it through a managed platform. On Pilae you pick Open WebUI from the catalog, choose a region, and click deploy. TLS, backups, patching and upgrades are handled for you from the first minute. You get self-hosting’s control without its ops.
Can Open WebUI connect to Ollama running on my own hardware?
Yes. Open WebUI speaks to Ollama and to any OpenAI-compatible API. With Pilae you can go further: run the whole deployment on your own GPU machines through our secure tunnel, so the interface, the models and the data all stay inside your own network.
Where does my chat data live?
In the region you choose: Zurich, Paris or Frankfurt, or on your own infrastructure. Conversations, uploaded documents and embeddings stay there, encrypted. For regulated workloads you can hold your own encryption keys.
How do Open WebUI updates and security patches work on Pilae?
We track upstream, prioritize CVEs, and ship updates locked to versions we have tested, rolled out with zero downtime and instant rollback. You stay current without watching the release feed.
Does Open WebUI support teams with multiple users, roles and SSO?
Yes. Open WebUI has multi-user support with role-based access control built in. For single sign-on, deploy Keycloak from the same catalog and put OIDC in front of it. Same console, same one-click deploys.
Can I move my Open WebUI deployment off Pilae later?
Yes, and that’s by design. Open WebUI is open source and your data exits with you in usable form: conversations, documents, configuration. No proprietary formats, no lock-in.
Runs well next to.
Build your stack from the same catalog. One console, one click each.
Workflow automation with 400+ integrations. Self-hosted, source-available.
Single sign-on, OIDC and SAML. The identity layer for the rest of the stack.
Metrics, logs and traces in one place. Dashboards pre-wired per pod.
Postgres-based development platform. Auth, storage, edge functions.
Open WebUI, without the ops.
Deploy it in minutes, keep your data in the country you choose, and leave whenever you want. Your data goes with you.