Ipyllm on Juampatronics's personal website

Use case for ipyllm: fixing issues of your django apps much faster

Sun, 01 Nov 2026 10:14:46 +0100

When you are hunting a performance issue or a bug, your model can only reason with your code and some logs or tracecalls you feed into the chat. It would be even more helpful, if your agent could access and even interact with the current state of your application. This is, in a nutshell, what ipyllm brings.

What is ipyllm? Link to heading

ipyllm is a open source IPython extension that seamlessly integrates Large Language Models (LLMs) into your live IPython environment. I have described it in previous posts. tldr: it is an agent with full access to your IPython environment, so it can reason about your actual state of the application and IPython history. Caution: it runs code in your environment too.n More on this later.

Bringing sessions to ipyllm

Fri, 30 Oct 2026 23:54:38 +0100

When doing data exploration it is also important to remember how you obtained the results. This is the main motivation behind this new version of ipyllm.

You can now start recording a session with %llm --save FILE.py. From that moment on, all your prompts will be saved as comments followed by the corresponding code produced by your agent in that file. %llm --save off stops the recording.

You can rerun a session and continue recording with %llm --resume FILE.py.

Supercharging IPython with ipyllm

Sun, 11 Oct 2026 23:57:42 +0200

ipyllm is a open source IPython extension that seamlessly integrates Large Language Models (LLMs) into your live IPython environment. Built on top of the versatile llm package, ipyllm extends the IPython REPL with agents that access your current environment to help you solve data analysis tasks, assist you with the exploration of an API or help you debug issues in your code.

What can ipyllm do for you? Link to heading

Effortless Data Exploration: Instead of explaining your dataset in text, ask %llm What columns in df have missing values? or %%llm Plot the rolling average of sales using matplotlib and run the code.. The model inspects the live variables, writes the visualization, and executes it in your session.
Contextual Debugging: When a script stalls or fails, drop ipyllm.embed() into your code. It works just like IPython.embed(), giving the assistant direct access to the function’s local scope and variables. You can ask: “What is holding result in a None state?” and let the model inspect local state on the fly.
Diagnostic Shell Integration: Run commands inside %%llm --run (such as pytest -q) to capture stdout, stderr, and exit codes instantly, letting the model explain failures and suggest fixes without manual log pasting.

Supercharging Data Analysis with Agent Skills Link to heading

Data analysis workflows often require specialized domain knowledge. ipyllm bridges this gap by supporting local Agent Skills.