Cursor Adds Inline Data Visualization to Agent Chats

AI-powered code editor Cursor has launched a new slash command, /visualize, allowing developers to render charts and diagrams directly inside their agent conversations. Available now within the Agents Window for all Cursor users at no additional cost, this feature targets quick data analysis tasks where building a full dashboard would be unnecessary. Instead of outputting raw Markdown tables or dense code blocks, the agent can now display visual representations of data inline.

To use the feature, developers type the command and instruct the agent to pull data from a local file, database query, log, or connected tool. This command integrates with Cursor's Model Context Protocol connections, which link the editor to external observability platforms and data sources like Datadog, Databricks, and Sentry. Practitioners can quickly plot database query results, inspect CSV data distributions, compare latency or error rates over time, and even sketch architecture or sequence diagrams during the planning phase of a project.

This update builds upon Cursor's existing Canvas feature, which allows agents to build persistent, interactive React-based dashboards in a side panel. While Canvas remains the preferred choice for complex interfaces that require ongoing editing and custom controls, the new inline visualization command is designed for rapid, exploratory debugging. It eliminates the friction of exporting raw data to external business intelligence tools or Jupyter notebooks just to perform a quick visual check. By keeping these visual insights directly within the chat thread, developers can maintain their flow state and make faster decisions during active debugging sessions.