Over the past year, AI copilots like Cursor and Windsurf have fueled a dramatic shift in software engineering workflows. And yet, many technical users in adjacent fields like data science and analytics have been unable to reap the rewards of this revolution.

It turns out that the existing tools are a poor match for analytical workloads. Beyond that Cursor and similar tools have very poor support for Jupyter notebooks, data science is a fundamentally different discipline from software engineering and we believe it requires a correspondingly different tool.

We're excited to announce the launch of Sphinx to fill this gap. Sphinx is a data-native AI copilot for Jupyter notebooks. Unlike existing copilots, Sphinx understands that data science often requires an experimental, exploratory approach.

Some highlights of Sphinx's capabilities: - A fully agentic loop, able to break tasks into discrete steps and implement them cell-by-cell - Able to create and read charts and graphs to guide its next steps - Fluent in working with common data connectors, like Snowflake, Databricks, and databases - Full MCP integration, and large outputs can even be processed programmatically as data directly in the notebook

To make it as easy as possible to use, Sphinx operates on ordinary Jupyter notebooks and is available as a simple VSCode extension with a generous free tier.

Excited to hear what you all think!