Repo

agentrewind

@amirshahzadhashmi7145by Amir Shahzad Hashmi · Independent

0

AgentRewind is a local tool for recording, inspecting, and replaying agent runs so you can debug why an agent made a bad decision without rerunning the whole thing from scratch. A lightweight Python SDK wraps LLM calls and tool calls with a decorator or context manager and records each step as a span: the prompt, model and parameters, response, too

Built withPython

See what other members are building and what each project is looking for.