An alternative to ChatGPT that stays in the Netherlands.
A chat environment for your staff, on AI models that run in a data centre in Ede.
ChatGPT is a good product. But when your staff paste customer files, contracts or patient data into it, you want to know where that goes and who can get at it. We set up a chat environment that runs in Ede, with the arrangements on your data in your contract.
The question is not whether your staff use AI.
They already do. The Dutch Data Protection Authority warned in 2024 and 2025 about data breaches caused by staff pasting personal data into chatbots: patient data at a GP practice, customer addresses at a telecom company, CVs and youth care files at a municipality. In most cases it was a free account, without the organisation knowing. The authority calls this shadow AI.
Banning it does not work, because the tool is too useful. What does work: a chat environment of the organisation itself, where staff can go without you having to wonder where the text ends up. That environment runs with us in a data centre in Ede, and what happens to your data is in your contract.
ChatGPT next to a chat environment in Ede.
An honest comparison on the points that matter to an organisation. ChatGPT wins on power and convenience; we win on where your data stays and who has a say over it.
Need the strongest model, or want to start tomorrow? Then ChatGPT Enterprise can be a sensible choice. If it is about files and conversations that had better not leave the building, an environment in Ede is the better fit.
ChatGPT and OpenAI are trademarks of OpenAI. Xyphen IT is not affiliated with or endorsed by OpenAI. Comparison based on public information as of 6 October 2026; check the current terms with the vendor.
A chat environment of your own organisation.
Not a login screen at an American company, but an environment we set up for you that runs in Ede.
What we do not deliver: the top model of the moment, web search and agents. What we do deliver is a fixed model version that only changes when we have agreed that with you.
This fits if you:
- Work with files, contracts or medical data and want to be able to explain where they are processed
- Are a municipality, healthcare provider, law firm or accountancy firm and want to replace shadow AI with an environment you control
- Mostly do text work: summarising, writing letters, questions about your own documents, transcribing conversations
- Would rather have a contract and a regular engineer than a separate account per employee
From intake to first conversation.
- IntakeWe discuss who will use the environment, for what, and which documents need to go in. If ChatGPT actually fits what you want better, we say so.
- Contract and arrangementsData processing agreement and the addendum on training and storage. Only then does real data come anywhere near the environment.
- SetupWe set up the chat environment, connect your documents and arrange who gets access. We walk through the first conversations with you.
- In useYour staff work with it; we keep the AI model and the hardware up to date. The model only changes by agreement, announced in a maintenance window.
How long this takes depends on your documents and your wishes. We tell you up front, not afterwards.
Frequently asked questions about an alternative to ChatGPT
Can this do what ChatGPT does?
No, not at everything. For summarising, letters, questions about your own documents and transcribing conversations, an open model of medium size does well. For complex reasoning, long documents, images and speech, ChatGPT remains stronger. In the intake we tell you honestly which tasks fit and which do not.
Does OpenAI train on my conversations?
That depends on your subscription. In the free version and Plus, training is on by default; you can switch it off yourself. In Business, Enterprise and Edu it is off by default. The problem with shadow AI is that staff often use the free version, with a personal account, and nobody checks that setting.
Can I not simply have my data processed in Europe at OpenAI?
For Enterprise and Edu you can, and for the API after approval. Not for the free version, Plus or Business. Bear in mind: OpenAI remains an American company, and the US CLOUD Act applies regardless of where your data is stored. With us there is no American company in the chain.
What was that preservation order in 2025?
In May 2025 a US court ordered OpenAI to preserve chats from the free version, Plus, Pro and Team, including chats users had deleted. The obligation for new chats ended in September 2025, but OpenAI still holds the data from that period. It shows that a court in another country can decide what happens to your conversations.
Can we start right away?
No. There is no trial account and no login screen you can use today. We start with an intake and then set up the environment per organisation. That is deliberate: the arrangements on your data should be in place before any data enters the environment.
Which AI model do you use?
Open models: models the maker releases so that others can run them on their own equipment. In the intake we pick the model that fits your task, and that version stays until we decide by agreement to change it. So you do not suddenly get a different model after an update.
Want to know whether an environment in Ede fits you?
Tell us who will work with it and for what. After an intake you receive a proposal with the model, the setup and the arrangements on your data.