Galdor – a Go LLM agent framework with built-in tracing and replay

Galdor is a Go-based framework for building and debugging LLM agents, featuring built-in tracing and replay capabilities. It is designed for developers who want to create reliable, observable AI workflows in Go without relying on Python-centric tools. Its interesting value lies in combining Go’s performance and concurrency with deep introspection, enabling precise debugging of agent behavior through full execution replay.