ipyllm

ipyllm is an IPython extension for discussing your live Python session with models supported by the llm package. Ask questions about your variables or REPL history, request plots, and have the model run code in the active IPython shell. It also works inside a live script: drop a debugging session with ipyllm.embed() to bring a coding assistant into that function's local scope.

Start with the installation guide, then see usage and commands.

Code runs with your permissions

Model-requested code executes in the live Python process. It is not sandboxed and can access files, the network, and session objects. Enable approval when you want to review code before execution; approval is not a security boundary.