I've seen this on occasion before, but the offering seems vague and confusing to me.
But asking to see if clarification is available:
* No API access at all. Is that correct?
* Usage limits aren't at all clearly defined, but their presence is suggested. Message per day/month, tokens per whatever limits?
* Content policy restrictions? They mention major models/providers.
* An actual list of models available? GPT, etc, are cool if that's all you need, but what about major/popular open source models?
* Encrypted/protected user content is mentioned, but do you allow verification by users that this is the case for compliance reasons? (This makes me think of providers that let you verify the nature of the secure computation yourself.)
This looks neat in theory, but there's nothing that goes into the exact technical nature of the offering. You may consider these suggestions for what might assist discerning technical users to have enough information about the service vs competitors that disclose all relevant information about the service up-front.
