Helicone

Helicone works as a proxy between your app and the model provider. You change one URL, and every request gets logged with its cost, latency and result, without extra code. From there you get dashboards, caching to cut repeat costs, rate limits and alerts. It also supports prompt versioning and experiments, and can route requests across providers with fallbacks. Helicone is open source and has a hosted version with a free tier and paid plans. It is an easy first step for teams that want visibility into their AI spending fast.