Skip to main content
Glama
stephaneberle9

vegalite-viewer

by stephaneberle9
OverviewSchemaSourceRelated ServersScoreDiscussionsAdmin
Local

Vega-Lite Viewer MCP Server

A Model Context Protocol (MCP) server that enables creating interactive data visualizations using the Vega-Lite grammar. Visualizations are rendered directly inside the chat using MCP Apps — no browser window required.

Usage

Prerequisites

This server requires uv. Install it via:

# Windows
winget install --id=astral-sh.uv -e

# macOS
brew install uv

# Linux
curl -LsSf https://astral.sh/uv/install.sh | sh

macOS note: The curl installer places uv in ~/.local/bin/ and updates your shell profile, but macOS GUI apps like Claude Desktop do not load shell startup files. Install via Homebrew to make uv visible to GUI apps.

See the uv installation guide for more options.

Quick Start

Add the following entry to your Claude Desktop configuration file (accessible via Settings... > Developer > Edit Config):

{
  "mcpServers": {
    "vegalite-viewer": {
      "command": "uv",
      "args": [
        "run",
        "--with-editable",
        "/path/to/mcp-server-vegalite-viewer",
        "mcp-server-vegalite-viewer"
      ]
    }
  }
}

Restart Claude Desktop to apply changes. The server is ready when vegalite-viewer appears in the list of connected MCP servers.

ℹ️ Note: Rendering visualizations inline in the chat requires a client that supports MCP Apps, such as Claude Desktop or Claude.ai.

CLI Reference

Flag

Description

--silent

Show only error messages

--debug

Enable detailed debug logging (also settable via VEGALITE_VIEWER_DEBUG=1)

MCP Tools and Prompt

Tools

Tool

Description

upload_data

Upload a JSON dataset and register it by name for later use in visualizations

visualize_data

Render a registered dataset as a Vega-Lite chart, displayed inline in the chat

Workflow: call upload_data first to register the dataset, then call visualize_data with a Vega-Lite specification to produce the chart. The same dataset can be visualized multiple times with different specs.

Prompt

Prompt

Description

Create a simple chart for a JSON dataset

Instructs the LLM to create a chart of a chosen type (bar, line, pie, etc.) for a provided JSON dataset

Example Prompts

Create a simple bar chart for the following JSON dataset:
[
  {"category": "Alpha",   "value": 4},
  {"category": "Bravo",   "value": 6},
  {"category": "Charlie", "value": 10},
  {"category": "Delta",   "value": 3},
  {"category": "Echo",    "value": 7},
  {"category": "Foxtrot", "value": 9}
]

Using with MCP Inspector

Create an mcp.json file:

{
  "mcpServers": {
    "vegalite-viewer": {
      "command": "uv",
      "args": [
        "run",
        "mcp-server-vegalite-viewer",
        "--debug"
      ]
    }
  }
}

Start the inspector from a terminal:

npx -y @modelcontextprotocol/inspector --config mcp.json --server vegalite-viewer

In your browser:

  • Click Connect to start the server

  • Go to Tools > List Tools to see the available tools

  • Find server logs under Server Notifications and in %TEMP%\mcp_server_vegalite_viewer.log (Windows) or ${TMPDIR:-/tmp}/mcp_server_vegalite_viewer.log (Linux/macOS)

Troubleshooting

Visualization not rendering inline

The client must support MCP Apps. In clients without MCP Apps support the tool still works — the Vega-Lite JSON spec is returned as text, which the LLM can describe or the user can paste into Vega Editor.

Server fails to start in Claude Desktop

Check the Claude Desktop logs:

  • Windows: %LOCALAPPDATA%\Claude\Logs\mcp-server-vegalite-viewer.log

  • macOS: ~/Library/Logs/Claude/mcp-server-vegalite-viewer.log

Or go to Settings > Developer, select vegalite-viewer and click Open Logs Folder.

Still having issues?

Run the server with --debug and open an issue on GitHub with the relevant log output.

Related MCP server: Data Visualization MCP Server

Contributing

See CONTRIBUTING.md for development setup, building the React app, code quality checks, and the release process.

Available Tools

2 tools
upload_dataA

A tool to upload and register a JSON dataset by name for use in subsequent visualizations. When to use this tool:

  • When the user asks to visualize data, call this tool first to register the dataset, then call visualize_data. How to use this tool:

  • Provide a short, descriptive name and the dataset as a list of JSON objects (records).

  • Each object should share a consistent set of keys (fields).

ParametersJSON Schema
NameRequiredDescriptionDefault
dataYes
nameYes

Output Schema

ParametersJSON Schema
NameRequiredDescription
resultYes

TDQS

A4.4/5.0
Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden. It discloses that the tool registers data for later use, implying persistence, but does not mention whether re-uploading with the same name overwrites, errors, or other side effects. Basic behavioral transparency exists, but lacks depth for a mutation tool.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is well-structured with clear sections for when and how to use. It is concise, with every sentence providing necessary guidance. No filler or redundant repetition of the schema.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's simplicity, the description covers the essential context: sequencing relative to visualize_data, data format guidelines, and the purpose. An output schema exists, so return values need not be described. Missing edge cases like name collisions or overwrite behavior, but for the core use case it is sufficiently complete.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 0%, so the description compensates by explaining that 'name' should be a short, descriptive label and 'data' should be a list of JSON objects with consistent keys. This adds meaningful semantics beyond the bare schema types, though it stops short of giving full examples or constraints.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's function: 'upload and register a JSON dataset by name for use in subsequent visualizations.' It uses specific verbs ('upload', 'register') and identifies the resource (JSON dataset) and its purpose, effectively distinguishing it from the sibling tool visualize_data.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicit usage instructions are given: 'When the user asks to visualize data, call this tool first to register the dataset, then call visualize_data.' It also provides how-to guidance on providing a name and data format, leaving no ambiguity about when and how to use the tool.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

visualize_dataA

A tool to render a Vega-Lite visualization of a registered dataset directly in the chat. When to use this tool:

  • When the dataset is complex or multi-dimensional and a visual representation would be more informative than text.

  • Not useful for single data points. How to use this tool:

  1. Upload the dataset first using the upload_data tool, then reference it by name here.

  2. Analyze the dataset's structure and fields to determine which fields map to which visual channels (x, y, color, size, etc.).

  3. Construct a Vega-Lite specification for the desired chart type. Consult the Vega-Lite documentation (https://vega.github.io/vega-lite/docs) and example gallery (https://vega.github.io/vega-lite/examples) for guidance.

  4. Ensure the spec is a valid instance of the Vega-Lite schema: https://vega.github.io/schema/vega-lite/v6.json. Do not include a "data" key — it is injected automatically from the registered dataset.

ParametersJSON Schema
NameRequiredDescriptionDefault
nameYes
specYes

Output Schema

ParametersJSON Schema
NameRequiredDescription
resultYes

TDQS

A4.6/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full burden of behavioral disclosure. It reveals that the 'data' key is injected automatically, requires a valid Vega-Lite spec, and mandates a prior upload step. However, it does not explicitly state that the tool is non-destructive or what happens with invalid specs, but the provided details are sufficient for safe use.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is well-structured with clear 'when to use' and numbered steps. It is longer than a typical one-liner, but every sentence provides necessary guidance for constructing valid Vega-Lite specs, and the structure aids comprehension.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's complexity and the presence of an output schema, the description covers prerequisites, usage steps, and critical constraints (e.g., doc link, schema validation, data injection). It gives an agent enough context to invoke the tool correctly, though it could mention error handling or result format.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The schema has no descriptions for the parameters (0% coverage), but the description thoroughly explains both: 'name' refers to a registered dataset, and 'spec' is a Vega-Lite schema instance with specific constraints (e.g., no 'data' key). This fully compensates for the missing schema info.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's purpose: 'render a Vega-Lite visualization of a registered dataset directly in the chat.' It uses a specific verb ('render') and resource ('Vega-Lite visualization'), and distinguishes it from the sibling tool 'upload_data' by focusing on visualization rather than data upload.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description explicitly provides when-to-use guidance ('When the dataset is complex or multi-dimensional...') and when-not-to-use ('Not useful for single data points.'). It also gives a clear prerequisite workflow: upload the dataset first using 'upload_data' and then reference it by name.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

  1. 2 tool updatesv0.1.0
    • First observedupload_data
    • First observedvisualize_data
Install Server
RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

TDQS

A4.5/5.0

Scored across 2 tools

Disambiguation5/5

Each tool has a clearly distinct purpose: upload_data registers datasets, while visualize_data renders them. The descriptions explicitly define when to use each, leaving no ambiguity.

Naming Consistency5/5

Both tools follow the same verb_noun pattern (upload_data, visualize_data), which is consistent and predictable.

Tool Count4/5

At 2 tools, the server is slightly under the typical 3-15 range, but the tools form a complete two-step workflow for the narrow purpose of Vega-Lite visualization, so the count is reasonable.

Completeness4/5

The upload-then-visualize flow covers the core workflow, but there are no management operations (e.g., listing or deleting datasets), which are minor gaps that agents can work around by re-uploading with a new name.

Maintenance

ActivityInactive
ResponsivenessNo issues

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Connectors

Related MCP Servers