Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@MCP Analyst Show me the distribution of ages in the dataset"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
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.
uv is required to run the MCP server.
Mac
brew install uvWindows
winget install --id=astral-sh.uv -eRelated MCP server: csv-mcp-server
To use the server in Claude you would need to update the Claude config
On MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json On Windows: %APPDATA%/Claude/claude_desktop_config.json
{
"mcpServers": {
"analyst": {
"command": "uvx",
"args": [
"mcp-analyst",
"--file_location",
"<replace_this_with_path_to_csv_or_parquet_files_on_your_machine>"
]
}
}
}The server supports the using globs in the file location. So for example if your directory data has more than one csv file then you can give file-location as <parent_path_to_data>/data/*.csv
Reads the data from the given file locations. Note that file_locations
can be a list of multiple files. However, all files must have the same schema
and the same columns. Executes the given polars sql query and returns the result.
Note that the polars sql query must use the table name as self to refer to the source data.
Inputs: file_locations (required), query (required), file_type.
Get the list of files that are source of data
Inputs: no named parameters.
Get the schema of a single data file from the given file location
Inputs: file_location (required), file_type.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 3 tools
Each tool has a distinct purpose: listing files, inspecting schema, and executing SQL queries. No overlap or ambiguity.
All names use snake_case with a verb_noun pattern. Two tools start with 'get_', while 'execute_polars_sql' uses 'execute_', which is a slight inconsistency but still predictable.
3 tools is slightly low but appropriate for a focused analysis server covering file discovery, schema inspection, and query execution. Not overly sparse.
The set covers the basic workflow of listing data sources, examining schema, and running SQL queries. Missing possibly result writing or data transformation, but core functionality is present.
Use your Mac, Windows or Linux computer from ChatGPT, Claude or Codex: files, commands, documents.
Your own computer's files from ChatGPT or Claude: folder rules, confirmations and undo.
Safe folder access for ChatGPT and Claude: read, write and search files, risky tools opt-in.
Give Claude only the Google Drive files you choose. Every action logged.
