What is Vizro MCP?
Vizro MCP is a server implementation based on the Model Context Protocol (MCP) designed specifically to enable LLM-powered creation and management of Vizro dashboards and charts. It provides a unified way to access validated design templates, connect to datasets, and produce high-quality, maintainable dashboard configurations with instant preview functionality.
How to Configure Vizro MCP
- Install MCP and Vizro MCP Server: Ensure MCP environment is set up on your machine, and install the Vizro MCP server package.
- Connect Data Sources: Prepare your data files—either local file paths or URLs to remote datasets.
- Start the Vizro MCP Server: Run the server locally to expose Vizro-specific functions and allow the LLM or MCP host application to connect.
- Security Consideration: For privacy, only run Vizro MCP on trusted, local environments, as there is currently no authentication or access control for the server.
How to Use Vizro MCP
- Launch your preferred LLM application or interface supporting MCP tool calls.
- Connect to the running Vizro MCP server.
- Interact with the agent by describing the dashboard or chart you wish to create. The LLM will orchestrate tool calls—such as fetching data info, planning dashboard layouts, or validating configs—through Vizro MCP.
- Receive instant feedback including live previews, validated code, and improvement suggestions at each design iteration.
- Use local or remote datasets by specifying their paths or URLs when prompted.
Key Features
- Unified Design Language: Ensures charts and dashboards conform to a common design standard and are easy to modify and maintain.
- Validated Configuration Output: All generated dashboard and chart code is checked, reducing errors and making editing easier.
- Live Preview: Instantly preview dashboard and chart configurations in your browser via direct integration with PyCafe.
- Flexible Data Access: Supports both local and remote datasets with minimal setup effort.
- Rich, Discoverable Toolset: Exposes a standardized set of tools for loading, analyzing, validating, and planning visualizations—all directly callable by LLMs.
Use Cases
- Interactive Dashboard Creation: Quickly build, iterate, and preview custom dashboards for presentations, reports, or internal analytics.
- Automated Chart Generation: Easily create charts based on data by simply describing the desired output to an LLM.
- Data Exploration and Analysis: Perform EDA (exploratory data analysis) directly through LLM interactions, leveraging structured analysis of uploaded or linked datasets.
- Template-Based Design: Use template-guided workflows to maintain visual and structural consistency across reporting assets.
FAQ
Q1: Can Vizro MCP be used remotely or in a cloud setup?
A: No, Vizro MCP is designed to run locally for privacy and security. There is currently no authentication mechanism for remote or hosted deployment.
Q2: What data formats does Vizro MCP support for charting and dashboard creation?
A: Vizro MCP supports CSV, JSON, HTML, Excel, OpenDocument Spreadsheet, and Parquet files as data sources.
Q3: How does live preview work in Vizro MCP?
A: When you validate a dashboard or chart configuration, the MCP server generates a preview link through PyCafe, which automatically opens in your web browser for real-time visualization.
Q4: Do I need to write or call Vizro MCP tools manually?
A: No, tool calls are managed by the LLM integrated with your MCP host. You only need to provide data and instructions in natural language.
Q5: Can I use Vizro MCP if my application does not support MCP tool calling?
A: Core functionalities will be limited if your host application does not support tool call operations.