Free trial
A free Hobby plan with monthly credits is available, so no separate trial period is needed to test the agent. Paid plans include more credits and advanced model access.
Learn moreRuncell
An AI agent that writes, runs and debugs Python cells inside JupyterLab
Intermediate
What is Runcell?
Runcell is a code modifier tool. A Jupyter-native AI agent that plans notebook steps, writes and executes Python, reads charts and tables, debugs errors and keeps project context across sessions, installed as a JupyterLab extension.
Runcell is an AI agent that operates inside JupyterLab rather than beside it. A question goes in, and the agent plans the notebook steps, writes the Python, executes the cells, reads the resulting tables and charts, and fixes errors as they appear. It works on the .ipynb files already open in JupyterLab, so the analysis stays in the environment where it runs. Installation is a pip command or the JupyterLab extension manager, followed by a restart and a sign in. There is no separate desktop app and no new editor to learn. The design point is the agent loop. Runcell inspects the notebook, data and existing code before choosing a step, then plans and executes, interprets cell outputs so the next decision rests on evidence instead of code alone, and carries the result forward into the following question. Context is kept across iterations and across sessions, so a multi-week project can be resumed by asking what has been done so far. A separate in-context assist mode handles smaller jobs: ask about a transformation or an error, and the agent reads the surrounding cells and applies a fix. It is built for analysts, data scientists, quantitative and risk researchers, and anyone whose work has to be reproducible rather than a single prompt. Against AI IDEs such as Cursor, the difference is notebook-native execution and output reading; against autocomplete, it is multi-step workflow work; against notebook chat, it is action on the notebook. GPT, Claude and Gemini models are included by plan, so no external API key is needed. The main constraint is scope. Runcell lives in Jupyter, which makes it a poor fit for teams working in scripts, IDEs or BI dashboards. Published material covers the free Hobby credits and paid plans but does not list exact credit allowances, per-seat limits or model tiers, so buyers should confirm cost before committing a team.
How do you use Runcell?
1Install the extension
Run pip install runcell, or add it through the JupyterLab extension manager, then restart JupyterLab so the agent loads.
2Create an account and sign in
Sign in from the extension. No OpenAI or Anthropic key is required because model access comes with the plan.
3Open an existing notebook
Load one of your own .ipynb files. Runcell leaves your workflow in place and adds a file tree, global search and git inside Jupyter.4Ask a question and let the agent plan
Describe the result you need. The agent inspects the data and code, breaks the task into steps, writes the Python and executes the cells.5Review outputs and iterate
Check the tables and charts produced, then ask the next question. Runcell reads the outputs and continues from the evidence rather than restarting.6Consult the documentation
The docs cover the agent loop and features in more depth, and the blog has tutorials on notebook conversion and pandas operations.
Pros and cons
Pros
- Executes cells and reads the outputs, so the next step follows actual results rather than code suggestionsAI
- Runs on existing .ipynb files inside JupyterLab with no separate editor or desktop appAI
- Cross-session memory of datasets and decisions supports long, multi-week analysesAI
- Bundled access to GPT, Claude and Gemini means no personal API key is neededAI
- Free Hobby plan with monthly credits lets a real project be tested before payingAI
Cons
- Very narrow scope: value depends entirely on working inside JupyterLab, with no use in script or IDE workflowsAI
- Homepage and FAQ give no credit amounts, seat limits or model tiers, so cost is hard to judge up frontAI
- Autonomous cell execution carries real risk on sensitive or production data without review checkpointsAI
- Model choice is tied to plan rather than fully user-controlled, which may matter for reproducibility or complianceAI
How much does Runcell cost?
Pricing
A free Hobby plan includes monthly credits for trying Runcell on a real project. Paid plans add more credits and access to advanced AI models for heavier notebook work. Exact credit amounts, tiers and prices are not stated on the homepage.
Learn moreSupport
Support is documented through the official docs and blog, which include tutorials and explainers. No dedicated help centre or contact channel is described on the homepage.
Learn moreIntegrations
Works inside JupyterLab on existing .ipynb files and adds git support in the notebook environment. Models such as GPT, Claude and Gemini are included by plan, so no external API key is required.
Features
Notebook-native agent in JupyterLab with multi-step planning, Python writing and cell execution, error diagnosis and recovery, and reading of tables, statistics, charts and image outputs. In-context assist explains or fixes cells using surrounding context, runnable experiments support method comparison, and cross-session memory keeps datasets and decisions connected. Also adds a file tree, global search and git inside Jupyter.
Learn moreFrequently asked questions about Runcell
How much does Runcell cost?
A free Hobby plan includes monthly credits for trying Runcell on a real project. Paid plans add more credits and access to advanced AI models for heavier notebook work. Exact credit amounts, tiers and prices are not stated on the homepage.Does Runcell offer a free trial?
A free Hobby plan with monthly credits is available, so no separate trial period is needed to test the agent. Paid plans include more credits and advanced model access.How do you use Runcell?
The walkthrough on this page covers 6 steps: 1. Install the extension 2. Create an account and sign in 3. Open an existing notebook 4. Ask a question and let the agent plan 5. Review outputs and iterate 6. Consult the documentation.What platforms does Runcell support?
Runcell is available on Web App, MacOS, Windows and Linux.What does Runcell integrate with?
Works inside JupyterLab on existing .ipynb files and adds git support in the notebook environment. Models such as GPT, Claude and Gemini are included by plan, so no external API key is required.What are the limitations of Runcell?
Very narrow scope: value depends entirely on working inside JupyterLab, with no use in script or IDE workflows. Homepage and FAQ give no credit amounts, seat limits or model tiers, so cost is hard to judge up front. Autonomous cell execution carries real risk on sensitive or production data without review checkpoints.
Status
StatusActive
Views0
Outbound clicks0
Added10/21/2026
Platforms
Web AppMacOSWindowsLinux
Pricing
FreeFreemiumSubscription
