Braintrust

Braintrust focuses on evaluation and experimentation for LLM applications. It helps you systematically test prompts, compare model outputs, and track improvements. The platform includes logging, a prompt playground, and tools for building evaluation datasets.