Choose which AI models touch your data.
Pick models by who made them and where they run, for the whole company. Or send work to a model you run yourself.
Most teams leave every model on and pay less. Regulated teams can switch whole countries off. Same setting, your call.
Choose by maker and by hosting Built
Switch off by country
Turn off models by the country that made them, or by where they run, for the whole company. A team or a bot cannot turn them back on.
Open models only, or your own
Allow only open-weight models, or send work to your own model first.
Nothing kept
Route only to providers whose terms say they keep nothing after the answer.
A record for your auditor
Every change to these settings is on a record that can only be added to.
Open models, run where you choose
Lower cost
Open-weight models (Llama, Mistral, Qwen, DeepSeek and others) can cost less than closed ones. Block any of them by the maker’s country with one switch.
Hosted where you pick
An open model hosted in the country you choose is not served by the company that made it.
Your own model Early
Connect your own model endpoint. We check it works, then send work to it first.
Hosted by us Coming
Open models on our own servers, at our own price.
Inside your own servers On request
Same product, your data centre
The whole workspace for your company alone, in your cloud account or your own data centre.
Set up with you
Planned with your team when you need it.
AI rules differ by country and sector; check yours with your counsel. No SOC 2 or ISO yet. Our AI laws research is a starting point.
Talk to us.
Your company, its size and the first job you would hand to an assistant. We answer by email, nothing else.
Tell us your company, its size and the first job you would hand over. A person answers.