CoML

An interactive coding assistant for data scientists, formerly MLCopilot, that works inside Jupyter Lab and Notebook with an OpenAI API key and a built in machine learning knowledge base.

CoML, formerly MLCopilot, is an interactive coding assistant for data scientists and machine learning developers, powered by large language models. It offers an out of the box natural language programming interface for data mining and machine learning tasks, integrated with Jupyter Lab and Jupyter Notebook. A built in knowledge base of machine learning backs up the answers during data work, which is what makes it useful beyond raw code generation.

Jupyter Lab is the recommended starting point. Before use, you export an OpenAI API key in the environment or set it in a dotenv file, then activate the extension in the notebook with a load command. The key has to be in place first, and the load command is what turns the assistant on inside the notebook.

A fix command repairs the cell above the current one, with optional details about what is wrong. That is the main interactive workflow shown in the README, and it is the kind of targeted action that fits a notebook loop well.

Current support is limited to Jupyter Lab and classical Jupyter notebook on Linux. The README notes that work continues on newer notebook versions and other environments, so the platform story is explicitly a work in progress.

CoML gives data mining and machine learning work a natural language prompt interface inside Jupyter, powered by an LLM and a built in ML knowledge base. Linux and a Jupyter environment are the current practical limits.

• Source 1 · Project repository
• Source 2 · README