Runcell

What is Runcell?

Runcell is an AI agent built for Jupyter notebooks, made by Kanaries Data Inc. You state the result you are after, and the agent works out the steps, writes the code, runs the cells and fixes errors along the way. You can also select a single cell, ask what is wrong or what to do next, and apply the generated fix in one click. For unfamiliar algorithms or libraries it can explain them with runnable examples, and domain experts such as researchers can ask for analyses or figures without writing the code themselves. Runcell installs as a JupyterLab extension with pip and also offers a Cloud Agent. Usage is metered in credits: chat and agent requests are charged from the tokens of the model you pick, converted at 30 credits per US dollar of model usage, while inline completions and similar helpers cost fixed small amounts. Pricing starts with a free Hobby plan; Pro costs $20/month, Pro+ $60/month, Ultra $200/month and Teams from $40/seat/month.

Use cases

• Automate multi-step data analysis in a notebook
• Fix a failing cell with one click
• Generate Python code from a plain-language goal
• Learn unfamiliar libraries with runnable examples
• Build publication-ready research figures
• Compare algorithms such as K-means and DBSCAN

Who it's for

• Data scientists
• Researchers
• Students learning data science
• Analytics teams

Why use Runcell

• The agent works inside the notebook, so code, outputs and fixes stay in one place instead of being pasted between a chat window and Jupyter.
• It reads the cells around your question, which makes error fixes and next-step suggestions relevant to your data.
• Token-based credits mean small requests cost little and you can see how model choice changes spend.
• Researchers can describe an analysis in domain terms and get working code without writing boilerplate.
• Paid plans include privacy mode and priority support.

Runcell features

• Autonomous agent that plans, writes and executes notebook cells
• Automatic error handling and recovery
• In-context help that reads surrounding cells and outputs
• One-click apply for generated fixes
• Explanations with live runnable examples
• Inline and simple code completion
• Image generation and visualization analysis
• Choice of models from Anthropic, OpenAI, Google and others
• Privacy mode on paid plans
• Team usage dashboard and seat management

Runcell integrations

• JupyterLab
• Stripe

How to get started with Runcell

1. Install Runcell in JupyterLab with pip install runcell, or open the Cloud Agent, and sign in.
2. Describe a goal in the agent, or select a cell and ask about it.
3. Runcell writes the code, runs the cells and reads the outputs, correcting errors as it goes.
4. Review the results and apply suggested fixes with one click.
5. Credits are deducted from your monthly allowance according to the model and tokens used.

Runcell limitations

• Credits run out: Hobby includes only 20 credits per month with limited model access
• Chat and agent cost varies with the model and the number of tokens used
• The pricing page notes the JupyterLab plugin service was temporarily unavailable, pointing users to Cloud Agent

Runcell pricing

Yes, Runcell has a free plan. Paid plans start at $20 per month, according to the official pricing page on 24 Oct 2026.

Runcell plans

• Hobby

  $0

  • 20 credits per month
  • Limited AI execution credits
  • Limited access to LLM models

• Pro

  $20/month

  • 500 credits per month
  • Access to all advanced models
  • Priority support
  • Privacy mode

• Pro+

  $60/month

  • Everything in Pro
  • 2,000 credits per month
  • 25% lower cost per credit than Pro
  • Priority access to new features
  • Priority support
  • Privacy mode

• Ultra

  $200/month

  • Everything in Pro+
  • 10,000 credits per month
  • Access to all advanced models
  • Priority support
  • Privacy mode

• Teams

  $40/seat/month

  starting price

  • Everything in Pro
  • 500 credits per seat per month
  • Centralized billing
  • Invite links
  • Team usage dashboard
  • Admin controls and seat management

Plans as listed on the official pricing page on 24 Oct 2026. Prices can change and may vary by country or exclude tax. Check the official pricing page.

Runcell alternatives

The main Runcell alternatives listed on Debuly are Jupyter AI, GitHub Copilot, Cursor, Julius AI, dbForge AI Assistant and Deepnote. The table compares pricing and platforms; order is A–Z.

1. 

   Cursor

   AI code editor and coding agent with cloud agents, Bugbot code review and multi-provider models, sold with Grok and X plans.

   CodingAutomation and agentsDeveloper platforms
   FreemiumVisit Cursor (official site)
2. 

   dbForge AI Assistant

   SQL helper built into Devart's dbForge database tools that writes, fixes, explains and tunes queries from plain-language requests.

   CodingProductivityDeveloper platforms
   FreemiumVisit dbForge AI Assistant (official site)
3. 

   Deepnote

   Collaborative AI notebook workspace for Python, SQL and no-code data analysis, with dashboards, scheduled runs and a built-in agent.

   CodingProductivityResearch and search
   FreemiumVisit Deepnote (official site)
4. 

   GitHub Copilot

   GitHub's AI coding assistant for code completion, chat and agents in VS Code, JetBrains, Visual Studio, the CLI and GitHub.com.

   CodingAutomation and agents
   FreemiumVisit GitHub Copilot (official site)
5. 

   Julius AI

   AI workspace for analysing data, researching, and building presentations, reports, websites, images and video from plain-language prompts.

   Research and searchProductivityDesign
   FreemiumVisit Julius AI (official site)
6. 

   Jupyter AI

   Open source JupyterLab extension that adds a chat panel where AI agents read, write and run notebooks with your approval.

   CodingDeveloper platformsResearch and search
   Open sourceVisit Jupyter AI (official site)

Runcell FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from runcell.dev · fetched 24 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.