Runcell

Runcell is an AI agent that works inside JupyterLab on existing .ipynb notebooks. Users describe a question or task, and Runcell inspects the notebook context, plans notebook steps, writes and executes Python cells, debugs errors, and reads tables, charts, statistics, and other outputs to inform the next step. The page states it can keep context across iterations and sessions, and that it works with existing notebooks. It is installed via pip install runcell or the JupyterLab extension manager. The page says the JupyterLab plugin service is temporarily unavailable and suggests using Cloud Agent in the meantime. The page states no API key is required and that access to models such as GPT, Claude, and Gemini is included based on plan.

Key features

• Autonomous agent that turns a question into executed multi-step notebook workflows
• In-context assistance that reads surrounding cells and outputs to apply fixes
• Explains concepts through runnable cells and side-by-side method experiments
• Writes and executes Python code in Jupyter notebooks
• Automatic cell execution and error diagnosis and debugging
• Reads tables, charts, statistics, and image outputs
• Works on existing .ipynb files in JupyterLab
• Keeps context of questions, outputs, and decisions across iterations
• Cross-session memory of dataset, decisions, and project state
• Access to multiple AI models such as GPT, Claude, and Gemini depending on plan

Best suited for

• Data analysts moving from business questions to reproducible notebook code
• Data scientists exploring data, testing methods, and comparing results
• Risk and quantitative finance analysis with diagnostics
• Researchers turning questions into code, figures, and reproducible evidence in Jupyter

Pricing

A free Hobby plan with monthly credits is stated; paid plans add more credits and access to advanced AI models. Specific prices are not stated on the website.