Jupyter AI is an open source Python extension for JupyterLab that connects AI agents via ACP, offering a native chat UI, file/command interactions, and multi-user collaboration. It includes MCP server integration and extensible entry points for custom agents.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
An open source extension that connects AI agents to computational notebooks in JupyterLab. It provides a native chat UI to collaborate with agents using the Agent Client Protocol (ACP). It supports agents reading/writing files, running terminal commands, and interacting with notebooks through a built-in MCP server. It includes a permission system for action approvals and supports multiple concurrent chats. It is designed to be flexible and extensible with custom MCP servers and an entry points API for building/registering agents.
How it works
Agents are automatically detected when dependencies are installed and can be used within JupyterLab via ACP. The extension leverages a built-in Jupyter MCP server to mediate interactions between agents and notebook content. It provides guardrails through a permission system to approve actions like writing files or executing commands. It allows dragging/dropping files or notebook cells as context and supports real-time collaboration with other connected users. Developers can add custom MCP servers and create their own AI personas using the entry points API, all while adhering to ACP and MCP standards.
Getting started
"Getting Started" link in the README is provided for installation, agent setup, and first chat. The repository README mentions quick links for getting started and user guides, but specific installation commands are not included in the truncated excerpt.
Recent releases
Latest releases include:
- v3.1.2 (2026-07-30): Maintenance and upkeep improvements; bump version.
- v3.1.1 (2026-07-27): Maintenance and upkeep improvements; bump jupyter.
- v3.1.0 (2026-07-23): Jupyter AI v3.1.0 with in-chat controls for a modern agentic editor in collaborative chats; AI personas expose controls including model.
- v3.1.0b2 (2026-07-15): Maintenance and upkeep improvements; prep Jupyter AI v3.1.
- v3.1.0b1 (2026-07-09): Bugs fixed; bump jupyter release pin.
Traction
Repository statistics: 4354 stars, 522 forks, 300 open issues. Language: Python. License: BSD-3-Clause. Created: 2023-02-09. Last push: 2026-07-30. Topics include acp, agents, jupyter, jupyterlab, jupyterlab-extension.
Behind the repo
Linked governance notes: Jupyter AI is under incubation as part of the JupyterLab organization.
Caveats
License is BSD-3-Clause. Open issues count is 300. Last push on 2026-07-30. Created 2023-02-09. No explicit age-related caveats beyond the release history; see open issues for ongoing maintenance.