- Tool type
- JupyterLab AI agent
- Made for
- Notebook-driven analysis
- Runs on
- JupyterLab extension
What it does
Runcell is a Jupyter-native AI agent for analysts, data scientists, and researchers working with notebook-based analysis, experimentation, or quantitative work. Give it a question or ask about a cell, transformation, result, or error. It inspects notebook context, plans steps, writes Python, executes cells, and reads outputs such as tables, statistics, charts, and images. It can diagnose execution errors and continue with follow-up notebook steps, leaving work in an inspectable .ipynb notebook.
Use it as an extension inside JupyterLab, not as a separate desktop application or a general-purpose web notebook runtime. Installation requires Python 3.10+ and JupyterLab 4.4.0+; classic Jupyter Notebook is not supported. The published plans allocate monthly AI credits: Hobby is free with 20 credits, while Pro starts at $20 per month when billed monthly. Sign-up and an internet connection are needed; Runcell says no separate model API key is required, and models are included according to plan. Outputs are generated by AI, so review code and results before relying on them.
What it can do
- Plans multi-step notebook workflows from a natural-language question
- Writes and executes Python cells
- Inspects existing notebook context and data
- Reads tables, statistics, charts, and image outputs
- Diagnoses execution errors and applies fixes
- Carries context and outputs across analysis iterations
Works with
- JupyterLab
- Existing .ipynb notebooks
Before you choose
- Runcell supports JupyterLab 4.4.0+ and does not support the classic Jupyter Notebook interface.
- AI usage is credit-limited: the free Hobby plan lists 20 credits per month; paid plans provide different monthly credit allowances and model access.
- Generated code and model outputs may be incorrect; review executions and results before acting on them.
- Paid privacy mode differs from free-plan data-use terms. The service processes requests with external model providers; local notebooks do not mean prompts remain on device.
Listed under
Get started
- Ensure Python 3.10+ and JupyterLab 4.4.0+ are installed.
- Install Runcell with `pip install runcell` and restart or launch JupyterLab.
- Open the Runcell extension from JupyterLab’s sidebar and sign up or sign in.
Plans and prices
Monthly billing: Hobby $0/20 credits, Pro $20/500, Pro+ $60/2,000 and Ultra $200/10,000 credits. Annual alternatives are $180, $600 and $2,000 respectively, billed yearly; allowances still renew monthly. Teams is $40/seat/month or $400/seat/year for 500 monthly credits per seat, starting at one seat. No separate model key is required. Confirm dollar currency and credit consumption for the selected models.
- 20 credits per month
- Limited AI execution credits
- Limited access to LLM models
- 500 credits per month
- Access to advanced models
- Priority support
- Privacy mode
- 500 credits per seat per month
- Centralized billing and team management
- Team usage dashboard
- Admin controls and seat management
Source: runcell.dev/pricing, checked 2 Nov 2026. Prices can change; check before you buy.
Sources for this overview
Product pages, documentation, and pricing checked 2 Nov 2026. Each source supports part of the overview; availability and prices can change.
- #1 Jupyter AI Agent - Runcell ↗www.runcell.dev
- About Runcell ↗www.runcell.dev
- Runcell Terms of Service ↗www.runcell.dev
- Runcell Pricing | Plans for AI-Powered Jupyter Assistance ↗www.runcell.dev
- #1 Jupyter AI Agent - Runcell ↗www.runcell.dev
Ratings and reviews
No ratings for Runcell yet.
What’s changed
Free accountSignals
Free accountData and security
Free accountCompare with similar tools
Free accountWhere this comes from
How we research- Company website
- runcell.dev: product, pricing and policy pages
- YC directory mirror
- Batch and company name
- Social accounts
- Accounts the company links, and its posts on X
- Human review
- Checked by a person
- knowledge_base
- manual_review
Last checked 2 Nov 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.