A private AI assistant for your whole organisation, on models you host, with prompts and documents that stay inside your network.
- Replacing
- ChatGPT Enterprise, by OpenAI
- With
- Open WebUI, LibreChat, Dify, vLLM, Ollama
- Runs on
- Your servers, or Pilae Cloud in Switzerland and the EU
- Migration
- Included in onboarding, at a fixed price
Why teams leave ChatGPT Enterprise
The documents people paste in
Contracts, patient files and draft accounts are what make an assistant useful. With models on your own machines, none of them leaves your network.
Your choice of model
Open-weight models on your GPUs, a hosted API where you accept the trade, or both side by side. Switching models is a configuration change, not a new contract.
Answers from your own files
Retrieval over your file share and wiki, with your permissions respected. The assistant quotes a document only to people who may already open it.
ChatGPT Enterprise vs Open WebUI operated by Pilae
| What is being compared | ChatGPT Enterprise | Pilae |
|---|---|---|
| Where prompts are processed | In OpenAI infrastructure, with data residency options for eligible Enterprise workspaces. | On your premises or on dedicated machines in any of 12 Pilae Cloud regions, six of them in Switzerland and the EU. Air-gapped on request. |
| Models | OpenAI models, updated by OpenAI. | Open-weight models you pick and pin, plus any hosted API you choose to connect. |
| Use of your data for training | Business data is not used for training by default, per OpenAI. | No training. The model runs on hardware you control, and chat history stays in your own database. |
| Sign-on and access | SAML SSO, SCIM provisioning and an admin console. | OIDC sign-on through Keycloak or Entra ID, with groups deciding who reaches which model. |
| Your own documents | File uploads, connectors to supported SaaS sources and custom GPTs. | Retrieval over your file share and wiki, indexed in pgvector beside the app, permissions respected. |
| Operation | Fully run by OpenAI as a SaaS product. | Run by the Pilae Agent and Pilae engineers: planned upgrades, verified backups, 60-second probes. |
| Pricing | Per seat, on an annual agreement. | On request, sized to your hardware and your team. No per-prompt charge. |
A private ChatGPT for your organisation
ChatGPT Enterprise is a capable product with serious enterprise controls. The question most organisations face is not whether it works. It is whether the documents people paste into it may leave the building at all.
With Pilae, the assistant runs where your data already lives: on your own servers, or on dedicated machines in any of 12 Pilae Cloud regions, six of them in Switzerland and the EU. Open WebUI gives your staff the chat experience they already know, with LibreChat or Dify where they fit better. The model behind it runs on hardware you control, served by vLLM or Ollama, so a prompt and its attachments stay inside your network. The full offer is on the private AI page, and data residency sets out where each piece runs.
Self-hosted AI, without the operations work
A self-hosted assistant fails in predictable ways: a model too large for the hardware, a database that does not survive real use, retrieval that ignores file permissions. We size the hardware before installing anything. The Pilae Agent then plans every upgrade, waits for your approval and records each change, and Pilae engineers are on call behind it.
Onboarding is fixed-price and includes the move. Pricing is on request: tell us about your team and we come back with a sizing and a quote.
Moving from ChatGPT Enterprise to Open WebUI
Size the hardware against the model
Which models your GPUs can serve, at what context length, for how many people at once. This is answered before anything is installed, and we say plainly where a local model is weaker.
Put it behind sign-on
Accounts from your directory through Keycloak or Entra ID. Nobody gets a new password, and a leaver loses access with everything else.
Connect your documents
Retrieval over the file share or the wiki, with permissions respected. Answers then cite the documents they come from.
Run both, then cancel the seats
The subscription stays live while we watch which tool people open. When the private one carries the load, you stop renewing.
The apps that replace ChatGPT Enterprise
Open WebUI
A chat interface over models you host, so prompts and the documents people paste into them never leave your network.
Replaces ChatGPT Team, Microsoft Copilot
LibreChat
One chat interface over every model your organisation allows, local or hosted, with the keys, the history and the choice of provider held on your side.
Replaces ChatGPT Team, Poe
Dify
A builder for AI apps, agents and retrieval workflows that runs next to your documents and your models instead of at a platform vendor.
Replaces OpenAI Assistants, Azure AI Studio
vLLM
An OpenAI-compatible inference server for open-weight models, run on GPUs in Switzerland, the EU or your own datacentre, so prompts and answers never reach a model vendor.
Replaces OpenAI API, Azure OpenAI
Ollama
A server that loads open-weight models on demand behind an OpenAI-compatible API, run on GPUs in Switzerland, the EU or your own datacentre so prompts never reach a model vendor.
Replaces OpenAI API, Azure OpenAI