They have a public API where you pay by usage, like most LLM companies. They also have a private API used by their apps that they sell under a subscription model (but with usage limits in exchange). Because those subscriptions are often cheaper than paying by usage, people reverse engineered the private API and put support for it in tools like OpenCode.

So the short of it is basically just that they want people to honor their ToS and not reverse engineer private interfaces for competing products.

As a customer you can of course have a different opinion on what they should do/what you want from them, which is where the drama comes from.

I think it’s also a good reminder that if your business is built on top of one of these AI companies, you are completely at their mercy. They can put limitations in their ToS breaking your product (e.g. it’s not allowed to use Claude to develop competing products), they can raise prices (effectively the same as the current drama), they can randomly lock you out or decide to not do business with you anymore, etc.

Not much of an implication because we knew this before. But a reminder of something to keep in mind especially knowing that none of these AI companies are profitable right now.