The Google Sheets MCP Server enables AI agents to interact with Google Sheets (and Google Docs) programmatically through a local Model Context Protocol server. It integrates with AI coding tools like Cursor, VS Code, and Claude Code.
Google Sheets capabilities:
List sheet tabs (sheets_list_tabs): Discover all tabs in a connected spreadsheet.
Read data (sheets_read_range): Read from any range using A1 notation (e.g., Users!A1:Z500); returns structured objects or raw 2D arrays, with control over value rendering (formatted, unformatted, or formula).
Get metadata (sheets_get_sheet): Retrieve spreadsheet title, URL, locale, and tab details (row/column counts).
Write data (sheets_write_range): Write a 2D array of values to a specified range for bulk inserts or updates.
Create a tab (sheets_create_tab): Add a new sheet tab to an existing spreadsheet.
Append a row (sheets_append_row): Append a JSON object as a new row, auto-aligning values with existing column headers.
Google Docs capabilities:
List, create, read, and get metadata for Google Docs.
Write or append text to a document.
Find and replace text within a document.
Key traits:
Runs entirely locally β API calls go directly from your machine to Google, with no third-party intermediaries.
Supports both Service Account (for teams/automation) and OAuth2 (for personal use) authentication.
Easy setup via npx google-mcp init wizard.
Allows AI agents to read, write, and query Google Sheets as a database, including listing sheets, reading ranges, writing data, creating tabs, and appending rows.
Click on "Install 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., "@Google Sheets MCP Server Show me the first 10 rows from the Sales sheet."
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.
Connect Google Sheets & Docs to Cursor, VS Code, Claude Code and AI agents β in 3 minutes.
Your spreadsheets and documents become data sources that AI agents can read, write, and query.
A Model Context Protocol (MCP) server that lets AI coding agents (Cursor, Copilot, Claude, Codex) interact with Google Sheets and Google Docs.
Once configured, you can tell your AI agent:
π Sheets:
"Read all users from the Users sheet"
"Add a new order row: name=Anton, total=150"
"Update the status column for row 23"
"List all sheets in my spreadsheet"π Docs:
"Create a new document titled 'Sprint Report'"
"Read the content of my meeting notes"
"Append a summary paragraph to the project doc"
"Find and replace all 'TODO' with 'DONE'"
"List all my Google Docs"Related MCP server: MCP Google Sheets Server
π·πΊ ΠΠ½ΡΡΡΡΠΊΡΠΈΡ Π½Π° ΡΡΡΡΠΊΠΎΠΌ: docs/quickstart-ru.md β ΠΏΠΎΡΠ°Π³ΠΎΠ²ΠΎ Π΄Π»Ρ Π½ΠΎΠ²ΠΈΡΠΊΠΎΠ².
Pick the auth method that fits your use case:
npx google-mcp initYou'll need:
Google Sheet URL β paste your sheet link
Service Account JSON key β download from Google Cloud Console
Use when you want the AI to act on your behalf β access your private sheets and docs without sharing them with a service account.
npx google-mcp init --auth oauthYou'll go through a browser-based OAuth2 flow. A refresh token is saved and auto-refreshed β you never need to re-login.
π When to use OAuth2 β β detailed guide and scenarios.
The init wizard will generate IDE configs for you automatically β just pick Project or System-wide when asked.
If you prefer manual setup, add to your IDE config:
{
"mcpServers": {
"google-mcp": {
"command": "npx",
"args": ["google-mcp"]
}
}
}{
"servers": {
"google-mcp": {
"type": "stdio",
"command": "npx",
"args": ["google-mcp"]
}
}
}{
"mcpServers": {
"google-mcp": {
"command": "npx",
"args": ["google-mcp"]
}
}
}{
"mcpServers": {
"google-mcp": {
"command": "npx",
"args": ["-y", "google-mcp"]
}
}
}More configs: examples/ β Cursor, VS Code, Claude, Codex.
That's it. Your AI agent now has access to your Google Sheets and Docs.
Command | Description |
| Interactive setup wizard (service account) |
| Setup with personal Google account (OAuth2) |
| Test connection + list sheets |
| Check OAuth2 refresh token health |
| List all sheets in the spreadsheet |
| Read data from a sheet |
| Create a new sheet tab |
| Append a row |
| Show current configuration |
| List Google Docs in your Drive |
| Read a Google Doc by ID |
| Create a new Google Doc |
π‘ Backward compatible:
npx google-sheet-mcpstill works as an alias fornpx google-mcp.
Once connected, AI agents get these tools:
Tool | What it does |
| List all sheet tabs with row/column counts |
| Read data from a range (returns objects with header keys) |
| Get spreadsheet metadata (title, URL, locale) |
| Write a 2D array of values to a range |
| Create a new sheet tab |
| Append a row (auto-aligns with headers) |
Tool | What it does |
| Create a new Google Doc |
| Read document content (plain text or structured) |
| Get document metadata (title, URL, revision, length) |
| Write/append text to a document |
| Find and replace text in a document |
| List Google Docs in your Drive |
Go to Google Cloud Console
Create a project (or use existing)
Enable Sheets API: APIs & Services β Library β "Google Sheets API" β Enable
Create Service Account: APIs & Services β Credentials β Create Credentials β Service Account
Give it a name β Create β Done
Click the service account β Keys β Add Key β Create New Key β JSON β Download
Share your sheet: Open your Google Sheet β Share β add the service account email (from the JSON) as Editor
π Detailed guide with screenshots: docs/setup-google.md
When to use OAuth2 instead of Service Account:
You want AI to access your personal sheets that you don't want to share with a service account
You're the only user and don't want to create a service account
Your sheets contain sensitive data that shouldn't be accessible via a shared key
Setup:
npx google-sheet-mcp init --auth oauthThe wizard will:
Ask for your OAuth2 Client ID and Client Secret (from Google Cloud Console)
Open a browser for you to grant access
Capture the authorization code automatically
Exchange it for a refresh token (stored locally)
How refresh tokens work:
The refresh token is stored in .google-sheet-mcp.json
Access tokens are auto-refreshed by googleapis β you never need to re-login
If a token becomes invalid, just run npx google-sheet-mcp init --auth oauth to replace it
Check token health:
npx google-sheet-mcp token-statusπ Key benefit: The AI agent acts as you β it accesses exactly the sheets you have access to. No need to share sheets with a service account email.
π Detailed guide: docs/setup-oauth2.md
Config is stored in .google-sheet-mcp.json (in your project or home directory):
{
"spreadsheetId": "1ABC...xyz",
"credentialsPath": "./credentials.json",
"sheets": ["Users", "Orders", "Payments"]
}Or use environment variables:
export GOOGLE_SPREADSHEET_ID="1ABC...xyz"
export GOOGLE_APPLICATION_CREDENTIALS="./credentials.json"google-sheet-mcp/
βββ src/
β βββ cli/ # CLI commands (init, test, list, read, append)
β βββ server/ # MCP stdio server + Google Sheets client
β βββ config/ # Config loader (.google-sheet-mcp.json + env)
βββ examples/ # MCP configs for Cursor, VS Code, Claude, Codex
βββ docs/ # Setup guides
βββ README.md
βββ package.jsonNo. Just use npx google-sheet-mcp init. No installation required.
Only "Editor" on the specific spreadsheet. Not on your entire Google Drive.
Yes. Use different config files per project, or set env vars per spreadsheet.
Yes. Share the sheet with the service account email (found in the JSON key).
No. The MCP server runs locally on your machine. Google Sheets API calls go directly from your machine to Google. No intermediate servers.
MIT
Append a row to a sheet. The row is a JSON object with column names as keys. The tool automatically reads the header row to align columns. Example: {"name": "Anton", "email": "anton@example.com"}
| Name | Required | Description | Default |
|---|---|---|---|
| row | Yes | Row data as JSON object. Keys = column headers, values = cell values. | |
| sheet | Yes | Sheet name (tab name). | |
| spreadsheet | No | Optional spreadsheet ID. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries the full burden. It discloses the key behavioral trait: automatic header row reading and column alignment. However, it does not mention side effects, permissions, or error scenarios, which would make it a 5.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two sentences plus an example. Every word adds value. Front-loaded with the core action. No redundancy.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's simplicity and absence of output schema, the description is mostly complete. It explains the core functionality and parameter semantics. It lacks potential error handling info but is sufficient for typical use.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so baseline is 3. The description adds value by providing an example and explaining the row parameter mapping (keys as column headers). It also clarifies the sheet parameter as 'tab name'.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb 'Append' and the resource 'a row to a sheet'. It explains the alignment via header row, which distinguishes it from siblings like write_range. The example further clarifies the purpose.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies use when adding a new row but lacks explicit when-not or comparison to alternatives like sheets_write_range. No direct guidance on when to choose this tool over siblings.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Create a new sheet tab in the spreadsheet.
| Name | Required | Description | Default |
|---|---|---|---|
| sheetName | Yes | Name of the new sheet tab. | |
| spreadsheet | No | Optional spreadsheet ID. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description should disclose side effects, prerequisites, and error conditions but only states the basic action.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
A single clear sentence, but could benefit from additional structure (e.g., specifying required vs optional parameters).
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Lacks details on tab name uniqueness, spreadsheet requirement, and output; for a simple creation tool, more context is expected.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema covers both parameters with descriptions, and the tool description adds no extra semantic value beyond what the schema provides.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb ('Create'), resource ('sheet tab'), and scope ('in the spreadsheet'), which distinguishes it from sibling tools like sheets_append_row or sheets_read_range.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No guidance on when to use this tool versus alternatives, such as when to specify the spreadsheet parameter or what happens if the tab name exists.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Get spreadsheet metadata: title, URL, locale, and all tabs with row/column counts.
| Name | Required | Description | Default |
|---|---|---|---|
| spreadsheet | No | Optional spreadsheet ID. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided; description does not disclose behavioral traits like read-only nature, permissions, or limits. Only states what is retrieved.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Single efficient sentence with no wasted words. Front-loaded with verb and resource.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Lists key metadata items. Without output schema, description is sufficient for agent to understand returns. Could mention if all sheets or active sheet, but 'all tabs' is clear.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100% with parameter description 'Optional spreadsheet ID.' Description adds no additional meaning beyond schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
Clearly states verb 'get' and resource 'spreadsheet metadata', listing specific items (title, URL, locale, tabs with counts). Distinguishes from siblings which are data operations.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No explicit guidance on when to use versus alternatives. While siblings are different, description doesn't provide context or exclusions.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
List all sheet tabs in the connected Google Spreadsheet. Use this to discover available sheets before reading or writing data.
| Name | Required | Description | Default |
|---|---|---|---|
| spreadsheet | No | Optional spreadsheet ID. Uses the configured default if omitted. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, and description lacks details beyond purpose. Does not disclose permissions, return format, or behavior when spreadsheet is missing. Minimal transparency for a simple read operation.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two sentences, no unnecessary words, front-loaded with action and resource.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple tool with 0 required params, the description is adequate. However, no mention of output format, which could help agents parse results.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Only one parameter 'spreadsheet' with description in schema. Description adds value by noting 'Uses the configured default if omitted', which is not in schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
Description clearly states 'List all sheet tabs' with specific verb and resource, and distinguishes from siblings by mentioning discovery before reading/writing.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Explicitly says 'Use this to discover available sheets before reading or writing data.' Provides clear context for when to use, though no explicit when-not-to or exclusions.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Read data from a Google Sheet range. Range format: 'SheetName!A1:Z100' or just 'SheetName' for the whole sheet. First row is treated as headers and data is returned as an array of objects.
| Name | Required | Description | Default |
|---|---|---|---|
| range | Yes | Range in A1 notation. Examples: 'Users!A1:Z500', 'Orders'. | |
| skip_header | No | If true, returns raw 2D array instead of objects with header keys. | |
| spreadsheet | No | Optional spreadsheet ID. | |
| value_render | No | How to render values. FORMATTED_VALUE = as seen in UI, UNFORMATTED_VALUE = raw numbers, FORMULA = show formulas. | FORMATTED_VALUE |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries full behavioral burden. It discloses that the first row is treated as headers and data is returned as an array of objects. It omits details like error handling, access requirements, or rate limits, but covers key expected behavior for a read tool.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is two sentences, front-loaded with the tool's purpose, and no unnecessary words. Every sentence provides essential information.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no output schema, the description explains the return format. Parameters are well-documented in the schema. The description covers the core functionality, though it could mention potential errors or limits. Overall, fairly complete for a read tool.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, so baseline is 3. The description adds value by explaining the range format with examples and clarifying header treatment. It does not add significant semantics beyond the schema for other parameters, so a 3 is appropriate.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool reads data from a Google Sheet range, specifies the range format, and explains the return format (array of objects with headers). This contrasts with sibling write tools like sheets_append_row and sheets_write_range, making its purpose distinct.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description does not explicitly state when to use this tool versus alternatives. It implies a read-only use case, but lacks 'when-not' guidance or reference to siblings for writes. The agent must infer usage from context.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Write values to a Google Sheet range. Values should be a 2D array: each inner array is a row. Example: [['name','age'],['Anton','30']]
| Name | Required | Description | Default |
|---|---|---|---|
| range | Yes | Range to write, e.g., 'Users!A1'. | |
| values | Yes | 2D array of values. First array = first row. | |
| spreadsheet | No | Optional spreadsheet ID. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries full burden. While it indicates mutation ('write'), it doesn't disclose whether the operation overwrites data, affects formatting, or has side effects like sheet creation. Behavior such as error handling or size limits is omitted.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is two sentences: the first states the purpose, and the second provides a concrete example. It is concise, front-loaded, and contains no extraneous information.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool has three parameters, no output schema, and no annotations, the description is adequate but incomplete. It covers the basic operation and value format but omits important context like range must exist, what happens on conflict, or success indication.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema already describes all three parameters with 100% coverage. The description adds value by explaining the expected structure of 'values' as a 2D array with an example, which clarifies beyond the schema's generic description. It does not elaborate on the 'spreadsheet' parameter but the example aids understanding.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states 'Write values to a Google Sheet range,' specifying the verb and resource. It distinguishes from siblings like sheets_read_range and sheets_append_row by indicating a write operation, and provides an example format for the values.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description does not provide guidance on when to use this tool versus alternatives (e.g., when to use write vs append). It lacks prerequisites, such as requiring the sheet to exist, and does not mention when the optional spreadsheet ID is necessary.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
v1.1.0sheets_append_rowsheets_create_tabsheets_get_sheetsheets_list_tabssheets_read_rangesheets_write_rangeEach tool targets a distinct operation: appending rows, creating tabs, getting metadata, listing tabs, reading ranges, and writing ranges. No two tools overlap in purpose.
All tools follow a consistent 'sheets_verb_noun' pattern (e.g., sheets_append_row, sheets_read_range), making it easy for agents to predict tool names.
With 6 tools, the set is well-scoped for a Google Sheets MCP server. It covers core operations without being bloated or overly minimal.
The tool surface covers basic CRUD for rows and tabs, and metadata retrieval, but lacks deletion operations (remove tab, delete row) and cell-level editing. Minor gaps for a typical spreadsheet workflow.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
Connect AI assistants to Google Sheets through controlled tools for reading and updating rows.
AI access to Quadratic spreadsheets: open files, run Python/SQL, query connected databases.
Query your Google Sheets as structured JSON: list sheets and tabs, read schemas, filter rows.
An agent-first office suite Claude & ChatGPT read and write over one MCP URL.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/we2go/google-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
