Test MCP server tools and chatbots at the ROOM level in the cloud. Resources spin up only for the life of the Room and are automatically cleaned up upon closure.

1. Install Meshagent

pip install "meshagent[all]"

2. Sign up and Authenticate

Follow instructions to sign up and log in:
https://docs.meshagent.com/cli/getting_started

3. Start a Room with Test MCP Server

meshagent service test --room=test --role=agent --image=meshagent/mcp-vizro:latest --env MESHAGENT_PORT=8001  --port="num=8001 path=/webhook liveness=/ type=meshagent.callable" --name=mcp-vizro-service-test
  • Starts a Meshagent Room named test with an MCP Vizro server agent.
  • The Room, MCP server, and tools will be shut down when the Room closes due to inactivity.

4. Start a Chatbot in the Room

meshagent chatbot join --room=test --agent-name=mcp-vizro --name=mcp-vizro --toolkit=mcp-vizro
  • Launches a chatbot in the test Room with access to the Vizro toolkit.
  • Multiple toolkits can be enabled in a Room with the same agent.
  • The command output provides a browser link to the Room.

5. Try it!

Visit the provided Room link in your browser.
Send messages to the agent and interact live with the MCP Server tools and chatbot.

Project Level Deployment: Persistent MCP Services

Deploy at the PROJECT level to automatically join every Room created in your Meshagent project. Great for production or persistent workflows. This removes the need to run the MCP server and chatbot locally — they auto-join on each Room creation.

1. Deploy the MCP Server Agent Service

meshagent service create --role=agent --image=meshagent/mcp-vizro:latest --env MESHAGENT_PORT=8001  --port="num=8001 path=/webhook liveness=/ type=meshagent.callable" --name=mcp-vizro-service

2. Deploy the Chatbot Service

meshagent service create --image="meshagent/cli:latest" --port="num=9001 path=/agent liveness=/ type=meshagent.callable participant_name=mcp-vizro-chatbot" --env="MESHAGENT_PORT=9001" --name="mcp-vizro-chatbot-service" --command="meshagent chatbot service --agent-name=mcp-vizro-chatbot --toolkit=mcp-vizro"
Now, every time a Room is created in that Meshagent project, the MCP server tools and chatbot are always available, with no extra setup required.

Tools Available

The MCP Vizro Server supplies tools and templates to help you create Vizro charts or dashboards step by step.
ToolShort Description
get_model_json_schemaGet the JSON schema for the specified Vizro model.
get_sample_data_infoIf user provides no data, use this tool to get sample data information.
get_vizro_chart_or_dashboard_planGet instructions for creating a Vizro chart or dashboard.
load_and_analyze_dataLoad data from various file formats into a pandas DataFrame and analyze it.
validate_chart_codeValidate the chart code created by the user and optionally open in PyCafe.
validate_model_configValidate Vizro model configuration.

Tool Details

get_model_json_schema

Get the JSON schema for the specified Vizro model.
  • Parameters:
    • model_name (string): Name of the Vizro model to get schema for (e.g., ‘Card’, ‘Dashboard’, ‘Page’).

get_sample_data_info

Obtain sample data details if no user data is provided.
Available datasets:
  • iris: Numerical + 1 categorical, good for scatter, histogram, boxplot, etc.
  • tips: Mix of numerical and categorical, bar, pie, etc.
  • stocks: Stock prices, line, scatter, time series.
  • gapminder: Demographics, line, scatter, maps, or many categories.
  • Parameters:
    • data_name (string): Name of the dataset.

get_vizro_chart_or_dashboard_plan

Get step-by-step instructions for creating a Vizro chart or dashboard.
Call FIRST when asked to create Vizro things.
  • Parameters:
    • user_plan (string)

load_and_analyze_data

Load data from files or URLs into a pandas DataFrame and analyze its structure. Supported formats: .csv, .json, .html, .htm, .xls, .xlsx, .ods, .parquet
  • Parameters:
    • path_or_url (string): Local file path or URL.

validate_chart_code

Validate created chart code and optionally open the PyCafe link.
  • Parameters:
    • chart_config (string): ChartPlan object with configuration.
    • data_info (string): Dataset metadata.
    • auto_open (boolean, optional): Open PyCafe in browser.

validate_model_config

Validate an entire Vizro dashboard configuration. If successful, returns Python code and, if remote, a py.cafe link.
  • Parameters:
    • dashboard_config (object): Vizro dashboard configuration (JSON/dict).
    • data_infos (array): List of DFMetaData objects.
    • auto_open (boolean, optional): Open PyCafe in browser.

Meshagent & MCP Resources

Visit Meshagent.com or read the docs for further details.
Find the official MCP Vizro server image on Docker Hub.
For more on MCP and Vizro: What is an MCP Server?
To deploy toolkits with Meshagent, start here.