Remember last week, when I said:
got an email yesterday: PlayAI is shutting down after being acquired by Meta. endpoints had been working sketchily for the past few weeks, and they finally announced that monthly plans will be cancelled automatically.
so, any project using their api will need to be reworked.
well, now i've learnt to own my own inference and not rely much on paid endpoints. rather work with self-hosted open source models, and even better if hosted locally without depending on a cloud provider.
I betrayed my own tought piece: I ended up using a different API endpoint. Why? Two reasons:
- I am trying to ship fast. It still is a partial technical debt that might bite me later on (like with PlayHT).
- The open-source models are not on the same level as the closed alternative (limited to English only, robotic voices).
I still prefer to deploy inference myself, but until some high-fidelity open-source model comes along, this will do.