I built a zero-dependency library to estimate LLM API costs before sending a request

If you build features on top of LLM APIs, you have probably been surprised by a bill at least once. Nothing is broken; you simply had no idea what a request would cost until the invoice arrived. So I