mcp-server-excel-sql 0.4.0

Let Claude query Excel files using SQL - no SQL knowledge required

mcp-server-excel-sql

Let Claude query your Excel files using SQL - no SQL knowledge required. Ask questions in plain English, Claude writes and executes the queries automatically.

What It Does

How it works:

1. Point the server at your Excel files
2. Ask Claude questions in plain English
3. Claude writes SQL queries automatically
4. Get instant answers from your data

Capabilities:

• Each Excel sheet becomes a queryable SQL table
• Join data across multiple spreadsheets
• Clean messy data with YAML transformation rules
• Deploy for teams with concurrent access
• Support for complex queries (aggregations, window functions, CTEs)

Should You Use This?

Great fit if you:

• Work with Excel files under 100MB
• Want data insights without SQL knowledge
• Need to join multiple spreadsheets
• Use AI assistants (Claude writes the SQL for you)
• Prototype before building ETL pipelines

Not the right tool if you:

• Have files over 100MB (use database import instead)
• Need to modify Excel files (read-only)
• Need formulas/macros/VBA (values only)
• Building production data warehouse (prototyping only)

Installation

Install uv:

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

That's it. No package installation needed - uvx runs the server on-demand.

Try It Now

Test with example data:

# Clone examples
git clone https://github.com/ivan-loh/mcp-excel.git
cd mcp-excel

# Generate example financial data
python examples/create_finance_examples.py

# Start server
uvx --from mcp-server-excel-sql mcp-excel --path examples

Server runs with 10 financial Excel files loaded.

Quick Start

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "finance-data": {
      "command": "uvx",
      "args": [
        "--from",
        "mcp-server-excel-sql",
        "mcp-excel",
        "--path",
        "/Users/YOUR-USERNAME/Documents/excel-data"
      ]
    }
  }
}

Customize:

1. Replace YOUR-USERNAME with your username (run whoami in terminal)
2. Update the path to your Excel files location
3. Restart Claude Desktop

Naming tip: Use descriptive names like finance-data, sales-reports, or excel for easy reference in conversations.

Command Line Testing

# Test with your files
uvx --from mcp-server-excel-sql mcp-excel --path /path/to/excel/files

# With auto-refresh
uvx --from mcp-server-excel-sql mcp-excel --path /path/to/files --watch

Common Use Cases

Financial Analysis

• Join budget vs actuals across quarters
• AR aging analysis across regions
• Revenue trending from monthly reports

Sales Reporting

• Combine sales data from multiple territories
• Product performance across time periods
• Customer segmentation from CRM exports

Operations

• Inventory reconciliation from warehouse exports
• Vendor comparison from procurement files
• Project tracking from multiple PM spreadsheets

Data Exploration

• Quick SQL access to ad-hoc Excel exports
• Testing data quality before building pipelines
• Prototyping analytics without database setup

How It Works

You ask Claude in plain English:

• "What's the total revenue by region?"
• "Show me customers with overdue invoices"
• "Compare Q1 budget vs actuals by department"

Claude automatically:

1. Writes the SQL query using the available tools
2. Executes the query against your Excel data
3. Returns formatted results

No SQL knowledge required - Claude handles query generation, table joins, and data formatting automatically.

Available Tools

The server exposes 4 MCP tools for working with Excel data:

tool_list_tables

• Lists all available tables loaded from Excel files
• Shows file path, sheet name, row count
• Call this first to discover your data

tool_get_schema

• Shows column names and types for a specific table
• Use after listing tables to understand structure

tool_query

• Execute SQL queries on your Excel data
• Read-only (no modifications allowed)
• Supports joins, aggregations, filtering

tool_refresh

• Reload data after Excel files have changed
• Automatic with --watch flag

Examples

The repository includes example Excel files with financial data for a Malaysian coffeehouse chain.

Generate examples:

python examples/create_finance_examples.py
uvx --from mcp-server-excel-sql mcp-excel --path examples --overrides examples/finance_overrides.yaml

Example queries:

-- Total debits in general ledger
SELECT SUM(COALESCE(debit, 0)) as total_debits
FROM "examples.general_ledger.entries";

-- Revenue by region
SELECT region, SUM(revenue) as total_revenue
FROM "examples.revenue_by_segment.revenue"
GROUP BY region
ORDER BY total_revenue DESC;

-- Budget variance analysis
SELECT
  department,
  budget_amount,
  actual_amount,
  (actual_amount - budget_amount) as variance,
  ROUND(((actual_amount - budget_amount) / budget_amount * 100), 2) as variance_pct
FROM "examples.budget_vs_actuals.data"
ORDER BY variance DESC;

Included files:

• General Ledger (MYR currency)
• Financial Statements
• Accounts Receivable Aging
• Revenue Analysis by Segment
• Budget vs Actuals
• Invoice Register
• Trial Balance
• Cash Flow Forecast

See examples/README.md for detailed query examples and usage patterns.

Understanding Table Names

Excel files are converted to tables with this naming pattern:

Format: <alias>.<filename>.<sheet>

Example:

• File: /data/sales/Q1-2024.xlsx
• Sheet: Summary
• Table name: sales.q12024.summary

Important: Table names contain dots and must be quoted in SQL queries:

-- Correct
SELECT * FROM "sales.q12024.summary"

-- Wrong (will fail)
SELECT * FROM sales.q12024.summary

Name cleaning:

• Converted to lowercase
• Spaces become underscores
• Special characters removed
• Only a-z, 0-9, _, $ allowed

System Views

Special tables for browsing loaded data:

<alias>.__files - File inventory

SELECT * FROM "sales.__files"

Shows: file paths, sheet count, total rows, modification time

<alias>.__tables - Table catalog

SELECT * FROM "sales.__tables"

Shows: table names, source file, sheet name, row count

Data Transformation

Excel files often have messy formatting. Use transformation rules to clean data.

What you can fix:

• Skip header/footer rows
• Combine multi-row headers
• Filter out total/summary rows
• Rename columns
• Set data types (dates, decimals, etc.)
• Pivot wide tables to long format

How it works:

1. Create YAML configuration file
2. Specify transformations per file/sheet
3. Load with --overrides config.yaml

Show transformation example

Create config.yaml:

sales.xlsx:
  sheet_overrides:
    Summary:
      skip_rows: 3                    # Skip header rows
      skip_footer: 2                  # Skip footer rows
      header_rows: 2                  # Combine multi-row headers
      drop_regex: "^Total:"           # Remove rows starting with "Total:"
      column_renames:
        "col_0": "region"             # Rename columns
      type_hints:
        amount: "DECIMAL(10,2)"       # Set column types
        date: "DATE"
      unpivot:                        # Pivot wide tables to long format
        id_vars: ["Region"]
        value_vars: ["Jan", "Feb", "Mar"]
        var_name: "Month"
        value_name: "Sales"

Run with overrides:

uvx --from mcp-server-excel-sql mcp-excel --path /data --overrides config.yaml

See examples/finance_overrides.yaml for complete real-world examples.

Modes:

• RAW (default): Loads Excel as-is with all columns as text, no headers
• ASSISTED: Applies transformation rules from YAML configuration

CLI Options

uvx --from mcp-server-excel-sql mcp-excel [OPTIONS]

Options:

• --path - Directory containing Excel files (default: current directory)
• --overrides - YAML configuration file for transformations
• --watch - Auto-refresh when files change
• --transport - Communication mode: stdio, streamable-http, sse (default: stdio)
• --host - Host for HTTP/SSE (default: 127.0.0.1)
• --port - Port for HTTP/SSE (default: 8000)
• --require-auth - Enable API key authentication (uses MCP_EXCEL_API_KEY env var)

Additional Documentation

Multi-user deployment, security, and development: See DEVELOPMENT.md for:

• Multi-user setup with authentication
• Security model and enforcement
• Architecture and design decisions
• Performance characteristics
• Testing and development workflow

Finance examples: See examples/README.md for detailed query examples and patterns.

License

MIT