Graham Dumpleton Releases AI-Built Wrapture Library

Python developer Graham Dumpleton, well-known for creating mod_wsgi, wrapt, and New Relic's Python agent, has introduced a new library called Wrapture. Released in late August 2026, the tool builds on the monkeypatching concepts of wrapt, expanding them to address both software testing and system tracing simultaneously. Wrapture allows developers to wrap any Python function or method, making it easy to monitor all incoming and outgoing data or to override return values entirely.

What makes Wrapture particularly notable for the AI industry is its development process. Dumpleton revealed that every line of code and documentation was generated by an AI assistant working under his strict guidance. He distinguished this from what he termed vibe coding, where developers blindly accept AI outputs. Instead, Dumpleton used his deep expertise in Python systems to architect the project, using the AI strictly as an execution tool to realize his specific design requirements.

For practitioners, Wrapture serves as a powerful alternative to the standard unittest.mock library and offers a non-intrusive way to implement tracing in existing codebases. The library features built-in OpenTelemetry support and includes a configuration-based mechanism that lets developers inject tracing into a Python project without modifying the underlying source code. For example, a developer can observe or alter the interaction when an order service calls a payment gateway charge method.

Though the project is only a few weeks old, it already demonstrates how experienced engineers can leverage AI to rapidly prototype and build robust system-level tools. By decoupling the architectural design from the manual writing of code, Wrapture showcases a highly disciplined approach to AI-assisted software engineering.