Google Sheets Data Editor MCP Server

Overview

The Google Sheets Data Editor MCP Server provides a secure, MCP-compliant bridge between large language models (or other model-driven tooling) and Google Sheets. It exposes a set of controllable tools for reading, writing, and bulk-editing spreadsheet data, and is designed to support workflows where models need safe, auditable access to tabular data without embedding credentials or spreadsheets directly into prompts.

By running a small server that holds a Google service account key and the target spreadsheet configuration, you can grant models limited, observable permissions to query and mutate sheet contents. Common benefits include bulk updates (batching multiple row edits in one operation), instantaneous sync of edits back into Sheets, and a simple tool surface that fits into MCP-based toolchains or orchestration layers.

Features

• MCP-compatible tool endpoints for model-driven workflows
• Read and write access to Google Sheets via a service account
• Bulk update operations for row-level batching and efficient writes
• Real-time syncing semantics (polling or push integration) to keep clients in sync
• Fine-grained configuration: spreadsheet ID, sheet ranges, read/write modes
• Secure deployment options: environment variable secrets, API token, HTTPS
• Audit-friendly request logging and optional rate limiting
• Lightweight: runs in Docker or as a standalone Node process

Installation / Configuration

1. Clone the repository and prepare credentials:

git clone https://github.com/xing5/mcp-google-sheets.git
cd mcp-google-sheets

2. Create a Google Cloud service account with the Sheets API enabled, download the JSON key, and share the target spreadsheet with the service account email.

3. Example .env file (create .env or pass env vars to Docker):

# .env
GOOGLE_APPLICATION_CREDENTIALS=/secrets/service-account.json
SPREADSHEET_ID=1AbcDeFGhIjKlmNOPq_rstUVwxyz1234567890
MCP_PORT=8080
MCP_API_KEY=replace-with-secure-token
SYNC_MODE=polling      # or "webhook" if configured
LOG_LEVEL=info

4. Run with Docker:

# mount credentials and pass env
docker run -it --rm \
  -p 8080:8080 \
  -v $(pwd)/service-account.json:/secrets/service-account.json:ro \
  -e SPREADSHEET_ID=your-spreadsheet-id \
  -e MCP_API_KEY=your-token \
  xing5/mcp-google-sheets:latest

5. Or run from source (Node.js example):

# Requires Node >= 16
npm install
export GOOGLE_APPLICATION_CREDENTIALS=/path/to/service-account.json
export SPREADSHEET_ID=your-spreadsheet-id
export MCP_API_KEY=your-token
npm start

Environment variables summary:

Available Resources

The server exposes a small set of MCP-style endpoints/tools designed for model orchestrators. Typical resources include:

• GET /mcp/info — server metadata and available tools
• GET /mcp/tools — list of tool descriptors (name, description, args)
• POST /mcp/run — execute a tool (read_range, update_row, bulk_update, watch)
• Webhook /events (optional) — event stream for spreadsheet change notifications

Example tool descriptors you can expect:

• read_range(range: string) — returns rows for an A1 range
• update_row(row_id: number, values: object) — updates one row by index/ID
• bulk_update(updates: array) — apply many update operations in a single transaction
• append_rows(values: array) — append rows to the sheet

GitHub repository: https://github.com/xing5/mcp-google-sheets

Use Cases

1. Data cleaning assistant

• Use a model to identify inconsistent entries in a column, produce a list of canonical values, then call bulk_update to rewrite affected cells in a single, auditable operation.

Example curl: read a range then bulk update

# read A1:C100
curl -H "Authorization: Bearer $MCP_API_KEY" \
  "http://localhost:8080/mcp/run" \
  -d '{"tool":"read_range","args":{"range":"Sheet1!A1:C100"}}'
# bulk update rows (example)
curl -H "Authorization: Bearer $MCP_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"tool":"bulk_update","args":{"updates":[{"row":5,"values":["fixed","value","2026-01-01"]}]}}' \
  http://localhost:8080/mcp/run

2. Automated report generation

• Periodically read specific ranges, feed them to an LLM for summarization, then write back a generated summary into a “Report” sheet using append