MCP Analyst
About
MCP Analyst is an MCP server that empowers claude to analyze local CSV or Parquet files.
Details
- Author
- unravel-team
- GitHub stars
- 18
- Downloads
- 241
- Categories
- Search
Jump to
- Works with local CSV and Parquet files
- Handles datasets larger than Claude’s context window
- Reduces cost by avoiding full file uploads
- Supports file globs for batch analysis
- Installed via uvx with a single command
Setting up with Highlight
This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:
- Download and install Highlight from highlightai.com/download
- Navigate to the plugins tab and select "Add Custom Plugin"
-
Configure the plugin with the settings below
Plugin Name
MCP AnalystCommand (node, npx, python, etc.)Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.
- Enable "Start Automatically" if you want the plugin to start when Highlight launches
From the repository
Install uv (required runtime) then add the server to your Claude desktop configuration file with the command uvx mcp-analyst --file_location <path_to_csv_or_parquet_files>. On macOS the config is at ~/Library/Application Support/Claude/claude_desktop_config.json; on Windows at %APPDATA%/Claude/claude_desktop_config.json.
Claude Desktop / Cursor
Paste into your MCP client config file to install this server.
{
"mcpServers": {
"mcp analyst": {
"analyst": {
"command": "uvx",
"args": [
"mcp-analyst",
"--file_location",
"<replace_this_with_path_to_csv_or_parquet_files_on_your_machine>"
]
}
}
}
}McpServers
{
"analyst": {
"command": "uvx",
"args": [
"mcp-analyst",
"--file_location",
"<replace_this_with_path_to_csv_or_parquet_files_on_your_machine>"
]
}
}MCP Analyst
MCP Analyst is an MCP server that empowers claude to analyze local CSV or Parquet files.
Use this server when your dataset size is bigger than the context window or you don't want to upload full file to optimize the cost.
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
