Cursor Lets Developers Visualize Data Inline During Agent Chats
The /visualize command turns data questions into inline charts and diagrams inside the Agents Window, no separate tool required.
- Cursor added /visualize, a slash command that renders charts and diagrams inline in the chat.
- Available now in the Agents Window for all Cursor users, no separate pricing.
- Extends the earlier Canvas feature, which renders persistent React artifacts beside the chat.
- Targets quick data analysis where a full dashboard would be overkill.
- Useful for query results, log files, and architecture or sequence diagrams during planning.
- Complements MCP integrations like Datadog, Databricks, and Sentry for observability workflows.
Cursor adds inline charts and diagrams to agent chat
Cursor has added a slash command that renders charts and diagrams inside agent conversations. In the Agents Window, users can type /visualize and direct the agent to data from a file, query, log, or connected tool. The result appears inline rather than as a long Markdown table or code block.
The command extends Cursor’s visual interface work into quick, conversational analysis. Cursor previously introduced Canvas, which lets agents build interactive dashboards in a side panel with React-based components for tables, charts, diagrams, and layout. /visualize brings a narrower version of that capability directly into the chat thread.
From tool output to chart
Agents can retrieve data from databases, local files, observability platforms, and MCP servers. MCP, or Model Context Protocol, gives an agent structured access to external tools and data sources. Those workflows often produce tables that preserve exact values but make trends, outliers, and relationships difficult to spot.
Inline visualizations suit several common development tasks:
- Plot query results from a database or MCP-connected service.
- Inspect distributions in a CSV, benchmark result, or debug file.
- Compare latency, error rates, or resource use across time windows.
- Draft architecture, dependency, or sequence diagrams during planning.
- Review incident data without moving it into a notebook or BI tool.
Developers can invoke the command after the relevant data enters the conversation and specify the desired view, such as a line chart, distribution, comparison, or diagram. A clear request should include the fields to plot, the grouping or time range, and any labels needed to interpret the result.
Where Canvas takes over
The Agents Window serves as Cursor’s hub for running agents across local repositories, worktrees, cloud environments, and remote SSH sessions. Canvas artifacts occupy durable panels alongside tools such as the terminal and browser, while /visualize keeps a quick result within the conversation.
| Task | Recommended tool |
|---|---|
| Inspect a query or metric once | /visualize |
| Build and refine a reusable dashboard | Canvas |
| Create a persistent interface with custom controls | Canvas |
| Sketch a diagram during an agent conversation | /visualize |
| Maintain repository-linked documentation | Canvas skill |
Canvas remains the stronger fit when a visualization needs continued editing, custom interactions, or reuse. The slash command reduces the setup required for exploratory work and short-lived debugging questions.
Rollout, cost, and open details
/visualize is available in the Agents Window for Cursor users and requires no separate setup. Cursor has not announced an additional fee for the command; its requests use the same agent workflow as the surrounding chat.
The initial announcement does not specify supported chart types, export formats, accessibility options, or whether inline visualizations can be saved independently from a conversation. Those details will determine how well the command fits reporting and documentation workflows. Its immediate use is clearer: developers can turn agent-accessible data into a visual check without transferring the results to another tool.
Comments
