Agentic Debugging with Time Travel: The Next Step for Coding Agents

Agentic Debugging with Time Travel: The Next Step for Coding Agents

What if you could rewind your program to see exactly where it went wrong?

In this episode of The Context, Mark Williamson, CTO at Undo, introduces agentic debugging—a revolutionary approach that combines time travel debugging with LLM-powered reasoning through the Model Context Protocol (MCP).

Mark demonstrates how Undo’s technology records every detail of a program’s execution, letting developers step backwards in time to pinpoint bugs, replay complex scenarios, and even collaborate with AI coding agents to interpret what happened.

From diagnosing cache corruption to asking an AI assistant when a zombie died in a Doom playthrough, this session explores how MCP integrations make debugging smarter, faster, and (surprisingly) more fun.

Find out more

Stay informed. Get the latest in your inbox.