Providers
Kagi uses a number of AI model providers, each with their own privacy policies, hosting country, etc. As a Teams plan administrator I might have to get individual providers approved for usage. Currently, I cannot prevent users from using models or providers which our risk assessment finds unacceptable for usage.
Even if I got all the current providers approved Kagi might add or change a providers which my users might use before we can assess them.
I note that Kagi are aware the differences between providers are of concern to users, as there is a note about differing privacy on one provider:
² When our Azure OpenAI Service API quota is saturated, the request is sent to the fallback provider: OpenAI's own API. OpenAI is currently required by court order to retain all chat logs. We are working with Microsoft to raise the Azure quota and minimize these fallbacks.
Models
Some models are much more expensive than others (Opus, o3 pro). And some usage modes consume a lot (Ki). It may be helpful to be able to only enable more cost-efficient models.
Impact
I don't think this would impact nor be useful to normal accounts.
I imagine this feature would be opt-in, with the default being unrestricted access to all providers/models, so no impact on existing workflows.
I would see team admins using this via settings. Issues I foresee:
- It might be confusing for users if they can't use a model listed in Kagi docs, so the UI might need to list them but disabled.
- Changing provider or models (e.g. chatgpt 4 -> 5) is a bit fiddly: as Kagi doesn't know why a model/provider was allowed/denied, they can't enable new options by default when mapping.
