About google-docs-mcp-server
google-docs-mcp-server lets Google Workspace users read and edit Google Docs through MCP. It handles nested tabs and converts document content to Markdown with heading structure.
Key features
- 01List nested document tabs
- 02Read tabs as structured Markdown
- 03Create documents and edit text
- 04Apply batch formatting and updates
- 05Search Drive for documents
Use cases
- 01Read every tab in a document
- 02Append or replace text in a chosen tab
- 03Apply formatting or table updates in batches
MCP support
It exposes Google Docs and Drive operations as tools for MCP clients, including tab-aware reading and editing. Authentication uses a Google service account with domain-wide delegation.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
uvx google-docs-mcp-serverSERVICE_ACCOUNT_PATHPath to the Google service account JSON key fileSUBJECT_EMAILEmail of the Workspace user to impersonate via domain-wide delegationLocal servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEGoogle Docs MCP Server
A Model Context Protocol server for Google Docs with full tab support — something most existing servers get wrong or skip entirely.
Uses Google service accounts with domain-wide delegation, so it works in Workspace orgs without OAuth consent screens.
What it does
| Tool | Description |
|---|---|
list_tabs | List all tabs (including nested) in a document |
read_document | Read one tab or all tabs as markdown with proper heading levels |
get_document_info | Lightweight metadata: title, ID, link, tab list |
create_document | Create a new doc with optional initial text |
append_text | Append text to the end of a specific tab |
insert_text | Insert text at a character index in a tab |
replace_text | Find and replace within a tab (or whole doc) |
batch_update | Full batchUpdate access — formatting, tables, images, styles |
list_documents | Search Drive for docs by name |
All tab-targeting tools accept tab_id or tab_title (case-insensitive). The batch_update tool auto-injects tabId into requests so you don't have to.
Why this exists
Google Docs has supported tabs since late 2024, but most MCP servers either:
- Ignore tabs entirely (only read the first tab)
- Don't use
includeTabsContent=True, so tab content is invisible - Don't handle nested tabs
This server handles all of that correctly and converts content to markdown with proper heading structure.
Install
pip install google-docs-mcp-server
Or run directly with uv:
uvx google-docs-mcp-server
Prerequisites
- Python 3.11+
- A Google Cloud service account with domain-wide delegation
Setup
1. Create a GCP service account
- Go to Google Cloud Console and create (or select) a project
- Enable the Google Docs API and Google Drive API
- Create a service account under IAM & Admin > Service Accounts
- Create a JSON key and download it
2. Enable domain-wide delegation
- In GCP, on the service account details page, enable Domain-wide Delegation and note the Client ID
- In Google Workspace Admin > Security > API Controls > Domain-wide Delegation
- Add the Client ID with these scopes:
https://www.googleapis.com/auth/documents https://www.googleapis.com/auth/drive.readonly
3. Configure your MCP client
Add to your MCP config (e.g. ~/.claude/mcp.json or .mcp.json):
{
"mcpServers": {
"google-docs": {
"command": "uvx",
"args": ["google-docs-mcp-server"],
"env": {
"SERVICE_ACCOUNT_PATH": "/path/to/your-service-account-key.json",
"SUBJECT_EMAIL": "[email protected]"
}
}
}
}
SUBJECT_EMAIL is the Workspace user the service account impersonates.
Environment variables
| Variable | Required | Description |
|---|---|---|
SERVICE_ACCOUNT_PATH | Yes | Path to the service account JSON key file |
SUBJECT_EMAIL | Yes | Email of the Workspace user to impersonate |
License
MIT
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.