UndoLog Documentation

Everything you need to understand, deploy, and extend UndoLog, the AI agent safe execution runtime.

Getting Started

Installation, quickstart, and core concepts to get you running fast.

Step-by-step guidance for tool annotation, compensations, approval gates, and integrations.

Complete API, configuration, and schema references.

Deep dives into the safety model, exactly-once semantics, and architectural decisions.

Architecture Decision Records: why we built UndoLog this way.

How to set up a development environment, write docs, and run tests.

Version migration guides and release notes.