Google Docs MCP Server
The Google Docs MCP server connects AI agents to your Google Documents. Agents can create documents, read and search their content, insert and replace text, apply formatting and styles, and manage document structure — letting assistants draft, edit, and maintain long-form documents in Google Docs directly from a conversation.
Install Google Docs
- Published by
- Official (vendor)
- Transport
- Remote (HTTP)
- Authentication
- OAuth
Claude Desktop, Claude Code and Cursor
Add this to the mcpServers object in your client's
config file, then restart the client.
{
"mcpServers": {
"google-docs": {
"type": "http",
"url": "https://docsmcp.googleapis.com/mcp/v1"
}
}
}
VS Code uses a different key — show that config
Identical entry, filed under servers rather than
mcpServers. Put it in
.vscode/mcp.json for one workspace.
{
"servers": {
"google-docs": {
"type": "http",
"url": "https://docsmcp.googleapis.com/mcp/v1"
}
}
}
Before it will answer
Google's Workspace MCP servers are NOT paste-and-go — this is the step people miss. You supply your own OAuth 2.0 client, created in the Google Cloud console as a Web application, and the redirect URI is client-specific (Claude uses https://claude.ai/api/mcp/auth_callback; Antigravity uses https://antigravity.google/oauth-callback). You must also enable TWO services in the project, not one: the product API and its separate MCP service. Here that is docs.googleapis.com and docsmcp.googleapis.com. Consent-screen scopes: drive.readonly, drive.file, documents.readonly and documents.