AI & LLM API Guides

Model routing, token costs, image and video generation, and what actually changes when you swap providers.

1 post in this category

We added 19 text-to-speech models. The hard part was billing them.

Nineteen TTS models now run through the same media queue as image and video. Here's why we bill on UTF-8 bytes instead of characters, and what broke.

One API key. 100+ endpoints. Yep, that's it.