Jupyter AI is an open-source extension that brings AI agents into JupyterLab, the notebook environment data scientists already live in. It adds a native chat interface where you collaborate with frontier coding agents directly alongside your notebooks, data and terminals.
Rather than locking you into one vendor’s assistant, Jupyter AI connects to agents through the open Agent Client Protocol (ACP): Claude, Codex, GitHub Copilot, Gemini, Goose, Kiro, Mistral Vibe and OpenCode are all supported, and agents are detected automatically once installed. Inside JupyterLab, agents can read and write files, run terminal commands and interact with notebooks through a built-in Jupyter MCP server, while a permission system requires your approval before an agent writes files or executes commands. You can run multiple concurrent chats, drag files or notebook cells into a conversation as context, and collaborate in real time with other users on the same server. Custom MCP servers can be attached for domain-specific tools, and developers can register their own AI personas through an entry-points API. It is governed as part of the wider Jupyter project, and its documentation covers installation, agent setup and building custom personas for different research domains. Jupyter AI is developed under the JupyterLab organization, is BSD-3-Clause licensed, has more than 4,400 stars on GitHub, and was under active development in 2026.
Free. Jupyter AI is open source under the BSD-3-Clause license; you bring your own agent subscriptions or API keys.
Data scientists, researchers and notebook users who want frontier coding agents working inside JupyterLab with their data and notebooks in context, without being tied to a single model vendor.