Connect Google Sheets to Claude: Sync cells and control permissions

If you need to connect Google Sheets to Claude to automate financial reporting, sync database exports, or manage workspace permissions, you need a Model Context Protocol (MCP) server. This server acts as the translation layer between Claude's JSON-RPC tool calls and the Google Workspace REST APIs. You can either build and maintain this OAuth infrastructure yourself, or use a managed integration platform like Truto to dynamically generate a secure, authenticated MCP server URL.

If your team uses ChatGPT, check out our guide on /connect-google-sheets-to-chatgpt-edit-data-and-manage-drive-files/ or explore our broader architectural overview on /connect-google-sheets-to-ai-agents-run-batch-updates-and-queries/.

Giving a Large Language Model (LLM) read and write access to a sprawling dataset like Google Sheets is an engineering challenge. You have to handle Google's strict OAuth 2.0 token lifecycles, manage the dichotomy between the Drive API and the Sheets API, and deal with complex A1 notation for cell targeting. Every time you need to expose a new capability, you have to update your server code, redeploy, and test the integration.

This guide breaks down exactly how to use Truto to generate a secure, managed MCP server for Google Sheets, connect it natively to Claude, and execute complex spreadsheet workflows using natural language.

The Engineering Reality of the Google Sheets API

A custom MCP server is a self-hosted integration layer. While the open MCP standard provides a predictable way for models to discover tools, the reality of implementing it against Google's APIs is painful. You are not just integrating a single endpoint - you are orchestrating multiple Google Workspace APIs that operate on entirely different design patterns.

If you decide to build a custom MCP server for Google Sheets, here are the specific integration challenges you will face:

The Drive vs. Sheets API Dichotomy An LLM does not inherently understand that a "spreadsheet" in Google's ecosystem is actually two different entities. If you want to read cell data, you use the Google Sheets API. If you want to copy the file, move it to a folder, or grant sharing permissions to a new user, you must use the Google Drive V3 API. A properly architected MCP server must expose tools from both APIs natively, allowing Claude to seamlessly transition from file management to data manipulation.

A1 Notation and Major Dimensions The Google Sheets API relies entirely on A1 notation (e.g., Sheet1!A1:D5) for reading and writing data. Unlike typical JSON REST APIs where you target an object by ID, Sheets requires precise grid coordinates. Furthermore, the API requires you to specify a majorDimension (ROWS or COLUMNS). If the LLM generates a malformed A1 string or mismatches the array structure with the chosen dimension, the API rejects the payload. Exposing these endpoints requires strictly defined JSON schemas that guide the LLM on exactly how to format these geometric requests.

Strict Quotas and Unforgiving Rate Limits Google enforces strict rate limits, often capping requests at 300 per minute per project, and 60 per minute per user per project for read requests. When you hit these limits, Google returns an HTTP 429 error. It is critical to understand that Truto does not retry, throttle, or apply backoff on rate limit errors. When the upstream API returns an HTTP 429, Truto passes that error directly to the caller. However, Truto normalizes the upstream rate limit information into standardized IETF headers (ratelimit-limit, ratelimit-remaining, ratelimit-reset). The caller - or the AI agent framework - is entirely responsible for inspecting these headers and executing the appropriate retry and backoff logic.

How to Create the Google Sheets MCP Server

Truto dynamically generates MCP tools based on the resources and documentation available for the Google Sheets integration. Each MCP server is scoped to a single authenticated Google account and secured via a cryptographic token.

You can create this server in two ways.

Method 1: Via the Truto UI

For administrators and internal tooling, the easiest path is the visual dashboard.

1. Navigate to the Integrated Accounts page in your Truto dashboard and select your connected Google Sheets instance.
2. Click the MCP Servers tab.
3. Click Create MCP Server.
4. Select your desired configuration. You can filter by specific methods (e.g., read-only) or specific tags to limit the LLM's access.
5. Click Create, and copy the generated MCP server URL (e.g., https://api.truto.one/mcp/abc123xyz...).

Method 2: Via the API

For developers building multi-tenant AI applications, you can generate MCP servers programmatically on behalf of your users.

Make an authenticated POST request to the /integrated-account/:id/mcp endpoint:

curl -X POST https://api.truto.one/integrated-account/{integrated_account_id}/mcp \
  -H "Authorization: Bearer YOUR_TRUTO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Claude Finance Assistant Server",
    "config": {
      "methods": ["read", "write"],
      "tags": ["googlesheets"]
    },
    "expires_at": "2026-12-31T23:59:59Z"
  }'

The API will validate the integration, ensure tools are available, and return a secure URL that you can immediately pass to your AI agent framework.

Connecting the MCP Server to Claude

Once you have your Truto MCP URL, you need to connect it to your LLM client. An MCP server is fully self-contained - the URL alone encodes the account scope and authentication.

Method 1: Via the Claude Desktop UI (or ChatGPT)

If you are using a consumer application like Claude Desktop or ChatGPT, you can add the server directly via the interface.

For Claude Desktop:

1. Open Claude and navigate to Settings -> Integrations -> Add MCP Server.
2. Paste your Truto MCP URL.
3. Click Add.

For ChatGPT (Requires Developer Mode):

1. Go to Settings -> Apps -> Advanced settings.
2. Enable Developer mode.
3. Under MCP servers / Custom connectors, add a new server.
4. Name it "Google Sheets" and paste the Truto MCP URL.
5. Click Save.

Method 2: Via Manual Configuration File (claude_desktop_config.json)

If you are running custom agents, utilizing the @modelcontextprotocol/server-sse transport, or configuring Claude Desktop manually, you can edit the configuration file directly.

Open your claude_desktop_config.json (located in ~/Library/Application Support/Claude/ on macOS or %APPDATA%\Claude\ on Windows) and add the SSE transport configuration:

{
  "mcpServers": {
    "google-sheets-truto": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-sse",
        "--url",
        "https://api.truto.one/mcp/YOUR_SECURE_TOKEN_HERE"
      ]
    }
  }
}

Restart Claude Desktop. The model will initialize the connection, execute a handshake, and dynamically pull the list of available Google Sheets tools.

Google Sheets Hero Tools

Truto automatically translates Google's API schema into MCP-compatible JSON schemas. Here are 7 high-leverage tools Claude can now access to orchestrate spreadsheet data.

get_single_google_sheets_spreadsheet_by_id

Retrieves the full metadata structure of a Google Sheet. This is critical for discovering the names of individual sheets (tabs) and determining the boundary ranges before attempting to read or write data.

"Claude, check the structure of spreadsheet ID '1BxiMVs0XRY...' and tell me the names of all the tabs inside it."

list_all_google_sheets_spreadsheets_values

Reads cell values from a specific A1 range. Empty trailing rows and columns are automatically omitted by the API, returning a clean 2D array of data.

"Read the data in the 'Q3 Revenue' tab from cells A1 to F20 in the main financial spreadsheet, and summarize the top three performing regions."

create_a_google_sheets_spreadsheets_value

Appends a new row of data to a spreadsheet. The Google API automatically detects the logical "table" within the specified range and appends the new values beneath the last populated row.

"Append a new row to the 'New Leads' tab in our CRM spreadsheet. The values should be 'Acme Corp', 'John Doe', and 'In Progress'."

update_a_google_sheets_spreadsheets_value_by_id

Overwrites existing cell data at a specific A1 coordinate. This is used for precise edits rather than appending new rows.

"Update cell D5 in the 'Q3 Revenue' sheet to '75000' to correct the European region total."

update_a_google_sheets_spreadsheets_values_batch_by_id

Updates multiple, distinct ranges in a single API call. This is essential for respecting Google's rate limits when the LLM needs to make dozens of scattered edits across a sheet.

"I need to update the statuses in the tracker. Change cell E2 to 'Approved', E7 to 'Rejected', and E12 to 'Pending'. Do this in a single batch operation."

google_sheets_files_copy

Leverages the Drive API to duplicate an existing spreadsheet. This is the foundation of automated client onboarding or report generation, allowing the agent to clone a template before filling it with data.

"Make a copy of the 'Monthly Report Template' spreadsheet and name the new file 'October 2026 Executive Summary'."

create_a_google_sheets_permission

Modifies the sharing and access controls of a Google Sheet. The LLM can use this to grant specific user emails read or write access to newly created reports.

"Grant 'reader' access to jane.doe@example.com for the new October summary spreadsheet we just generated."

For the complete inventory of available Google Sheets tools, including metadata search, Drive label management, and data filter operations, visit the Google Sheets integration page.

Workflows in Action

Giving Claude raw API tools is powerful, but the real value lies in multi-step orchestration. Here is how an AI agent strings these tools together to execute complex operations.

Scenario 1: Automated Financial Reporting and Batch Updates

An operations manager needs to reconcile a list of recent expenses, update the main ledger, and correct specific faulty entries across multiple tabs.

"Claude, pull the recent expense data from the 'Raw Imports' tab in the master ledger. Calculate the department totals, append the new totals to the 'Monthly Summary' tab, and correct the misspelled vendor name in cell C14 of the imports tab."

Execution Steps:

1. list_all_google_sheets_spreadsheets_values: Claude reads the data from the Raw Imports!A1:G100 range to analyze the recent expenses.
2. Internal Processing: Claude processes the array, calculating the sums for each department and identifying the misspelled vendor.
3. create_a_google_sheets_spreadsheets_value: Claude appends the freshly calculated department totals to the bottom of the Monthly Summary table.
4. update_a_google_sheets_spreadsheets_value_by_id: Claude targets cell Raw Imports!C14 and overwrites the misspelled vendor name with the correct spelling.

sequenceDiagram
    participant User as User
    participant Claude as Claude Desktop
    participant Truto as Truto MCP
    participant Google as "Google Sheets API"
    User->>Claude: "Analyze expenses, append totals, fix typo"
    Claude->>Truto: Call list_all_google_sheets_spreadsheets_values
    Truto->>Google: GET /v4/spreadsheets/{id}/values/Raw Imports!A1:G100
    Google-->>Truto: Return 2D array
    Truto-->>Claude: Return cell data
    Claude->>Truto: Call create_a_google_sheets_spreadsheets_value
    Truto->>Google: POST /v4/spreadsheets/{id}/values/Monthly Summary:append
    Google-->>Truto: Return append confirmation
    Truto-->>Claude: Return success
    Claude->>Truto: Call update_a_google_sheets_spreadsheets_value_by_id
    Truto->>Google: PUT /v4/spreadsheets/{id}/values/Raw Imports!C14
    Google-->>Truto: Return update confirmation
    Truto-->>Claude: Return success
    Claude-->>User: "Totals appended and typo corrected."

Scenario 2: Template Duplication and Client Onboarding

An account manager needs to spin up a new tracking dashboard for a newly signed client, populate it with initial data, and share it with the client's point of contact.

"Claude, a new client 'Stark Industries' just signed. Clone our standard 'Client Onboarding Tracker' template, add their company name to the header in cell A1, and share the new file with tony@stark.com so he can edit it."

Execution Steps:

1. google_sheets_files_copy: Claude calls the Drive API to duplicate the template file, capturing the new id returned in the response payload.
2. update_a_google_sheets_spreadsheets_value_by_id: Using the new file ID, Claude writes "Stark Industries Onboarding" into the Sheet1!A1 range of the cloned spreadsheet.
3. create_a_google_sheets_permission: Claude uses the Drive API to create a new permission record on the new file ID, setting role to writer, type to user, and emailAddress to the client's email.

flowchart TD
    A["Claude Agent"] -->|"google_sheets_files_copy"| B["Duplicate Template<br>(Drive API)"]
    B --> C{"Extract New File ID"}
    C -->|"update...value_by_id"| D["Write Client Name<br>to A1 (Sheets API)"]
    C -->|"create...permission"| E["Grant Edit Access<br>to Email (Drive API)"]
    D --> F["Workflow Complete"]
    E --> F

Security and Access Control

Giving an AI agent write access to your corporate Google Drive requires strict governance. Truto provides multiple layers of control when configuring your MCP server tokens:

• Method filtering: You can restrict a server strictly to read operations. Setting the config to methods: ["read"] ensures the LLM can only call list and get operations, physically blocking it from writing data, copying files, or changing permissions.
• Tag filtering: You can scope the server down to specific functional areas using tags. If you only want the LLM to access Drive metadata and not cell data, you can filter the generated tools accordingly.
• Require API token auth: By setting require_api_token_auth: true, the URL alone is no longer sufficient. The MCP client must also pass a valid Truto API token in the Authorization header, ensuring only authenticated internal users can execute tools via the agent.
• Time-to-live (TTL) expiration: You can enforce temporary access by setting an expires_at timestamp. Once the timestamp is reached, the underlying Key-Value storage automatically evicts the token, instantly revoking the LLM's access to Google Sheets without manual intervention.

Summary

Connecting Claude to Google Sheets via MCP transforms a static LLM into an automated data analyst and workspace administrator. By abstracting the complexities of Google's dual Drive/Sheets APIs, handling the idiosyncrasies of A1 notation, and securely managing OAuth lifecycles, Truto allows engineering teams to focus on agent orchestration rather than integration maintenance.

Stop writing custom scripts to ferry data between your LLM and your spreadsheets. Rely on a managed MCP infrastructure to keep your workflows secure, scalable, and resilient against API drift.