Excel Analyser MCP

What it does

Excel Analyser MCP connects AI agents and LLM clients to local or remote Excel (.xlsx), CSV (.csv), and JSON (.json) files. It exposes a set of tools that let the AI read file metadata, preview rows, filter by column, query by field value, and paginate through large datasets in chunks — without ever loading the entire file into memory. A ready-to-use hosted HTTP endpoint (Railway) is available for instant zero-install access.

Tools

• read_excel — Read an Excel or CSV file, returning a preview (first 100 rows) and metadata for large files, or full data for small files.
• get_chunk — Fetch a configurable batch of rows from a CSV or Excel file starting at any offset.
• read_json — Stream-read a JSON file and return a preview with field names and total entry count.
• query_json — Memory-efficient field-value search across multi-gigabyte JSON files using contains/equals/startsWith/endsWith operators.
• get_json_chunk — Iterate through a large JSON array in configurable chunks for sequential processing.

Installation

Claude Desktop (stdio — recommended for local files):

{
  "mcpServers": {
    "Excel Analyser MCP": {
      "command": "npx",
      "args": ["-y", "excel-analyser-mcp"]
    }
  }
}

Claude Desktop / any HTTP client (hosted — no install):

{
  "mcpServers": {
    "Excel Analyser MCP": {
      "type": "http",
      "url": "https://web-production-64851.up.railway.app/mcp"
    }
  }
}

Self-hosted HTTP:

npx excel-analyser-mcp streamableHttp

Then point your client at http://localhost:8080/mcp.

Supported hosts

Confirmed in README: Claude Desktop, Cursor AI, VS Code Copilot. HTTP transport also works with any MCP client that supports streamable HTTP.