Documentation

Enable the API for each service you plan to use, in the same Google Cloud project as your OAuth client. Each link opens the Cloud Console enable flow for whichever project is currently selected, so check the project picker first. To enable them all from the CLI instead, run the gcloud command below with your project ID, dropping any services you will not use. Calls to a service whose API is disabled fail with a 403 SERVICE_DISABLED error. Google Chat also needs a Chat app configured on the Chat API Configuration tab before it accepts requests.

gcloud services enable \
  gmail.googleapis.com \
  drive.googleapis.com \
  calendar-json.googleapis.com \
  docs.googleapis.com \
  sheets.googleapis.com \
  slides.googleapis.com \
  forms.googleapis.com \
  tasks.googleapis.com \
  people.googleapis.com \
  chat.googleapis.com \
  customsearch.googleapis.com \
  script.googleapis.com \
  --project=PROJECT_ID

Enable with MCP_ENABLE_OAUTH21=true and run HTTP transport (--transport streamable-http). This is the recommended mode for modern MCP clients and bearer-token auth.

Default mode when OAuth 2.1 is not enabled. Use this for older clients that do not support OAuth 2.1. --single-user cannot be combined with OAuth 2.1.

Container-friendly mode with no filesystem writes. Requires OAuth 2.1 and bearer tokens on requests.

Use when your app handles OAuth externally. The server skips local OAuth flow and validates incoming bearer tokens.

Select proxy storage backend for OAuth state in distributed deployments.

Authenticate as a service account and impersonate domain users. When a tool call supplies user_google_email, it is used as the per-request impersonation subject, so one server instance can act for multiple domain users; USER_GOOGLE_EMAIL is the fallback subject.

Customize the OAuth 2.1 consent page shown to users during authorization.

Run with --transport streamable-http for Claude Code, VS Code MCP, MCP Inspector, and OAuth 2.1 support.

Use only for clients with incomplete MCP support. Not recommended for modern multi-user flows.

Directly list and invoke tools from the command line without running the full MCP server process.

Three cumulative tiers for simplified deployments: core, extended, complete.

--read-only requests readonly scopes and hides write tools. --permissions sets a per-service access level and implicitly selects which services load — only the services you list are imported. Levels are cumulative: each one includes every level to its left. Mutually exclusive with --read-only and --tools; combine with --tool-tier to narrow further within the selected services.

Use external URL overrides when running behind a reverse proxy so generated OAuth endpoints match public domains.

Downloaded Gmail/Drive attachments are stored in persistent local storage and exposed by URL in HTTP mode.

Distributed tracing via the otel extra (pip install "workspace-mcp[otel]", included in the Docker image). Tracing is fully disabled unless an OTLP endpoint is configured; spans export only to the collector you specify. Standard OTLP variables for headers, TLS, sampling, and batching are honored.

Server logs default to INFO. DEBUG includes companion lines that may contain user text, so enable it only when needed and treat its output as sensitive.

Lists calendars accessible to the authenticated user, with each calendar's time zone.

Retrieves events for a calendar and time range.

Consolidated create, update, and delete event operations.

Create secondary calendars and manage who a calendar is shared with. Replaces create_calendar.

Query free/busy information across one or more calendars.

Create, list, update, or delete Out of Office events.

Create, list, update, or delete Focus Time events.

Search Drive files and folders.

Read file content with export support. Images come back as image content, SVGs as markup with any embedded rasters attached. Long text comes back a page at a time (40,000 characters by default) with the offset to continue from.

Download Drive files via local path or HTTP attachment URL, or inline as an embedded resource in stateless mode unless a GCS attachment bucket is configured.

Create files from inline content or source URLs, or create a shortcut to an existing file or folder.

Create folders in My Drive or shared drives.

Import files (Markdown, DOCX, HTML, etc.) into Google Docs.

Import presentation files (PPTX, PPT, ODP) as Google Slides.

Import spreadsheet files (XLSX, CSV, TSV, etc.) as Google Sheets.

Generate shareable links for files.

List folder contents or shared drives.

Copy existing Drive files with optional renaming.

Update metadata, move files, or replace/append to Google Apps content in place.

Grant, update, revoke permissions, and transfer ownership.

Configure link sharing and file-level sharing.

Get file metadata, parents, and detailed permission entries.

Check whether a file is publicly accessible.

Search mail using Gmail query operators.

Retrieve message content. With full=True, returns the complete untruncated message via file/URL (or inline when no file storage is available).

Batch fetch message content.

Send email messages (supports replies, forwards, and attachments). When thread_id is set, threading headers are derived from the thread automatically. Sends are never retried: a timeout or server error reports that the message may already have been delivered, so check Sent before sending again.

Retrieve full thread content.

Download Gmail attachments to persistent attachment storage.

Add/remove labels on a message.

List account labels.

List configured Gmail filters.

Create/update/delete labels.

Create, update, and delete Gmail filters, or apply one to existing mail.

Create Gmail drafts, including a forward of an existing message saved for review.

Batch fetch thread content.

Batch label modifications on messages or whole conversations.

Legacy OAuth 2.0 auth helper (disabled with OAuth 2.1).

Extract plain text document content, whole or a single heading section.

Create new documents.

Modify document text with formatting and links.

Search docs by name.

Find and replace text.

List docs in a Drive folder.

Insert tables, lists, and page breaks.

Apply heading and paragraph styles, including nested lists and per-edge borders.

Export a document as formatted Markdown.

Export a doc to PDF.

List all document comments.

Create, reply to, resolve, update, or delete comments. Update and delete apply to top-level comments you wrote.

Insert images from Drive or URLs.

Update headers and footers.

Execute multiple doc operations in one call, including insert_person to add a person chip by email and insert_date to add a date chip (YYYY-MM-DD or an ISO 8601 timestamp, with optional date format, time format and timezone).

Inspect structural document elements. Both modes also report paragraph layout stats: empty-paragraph counts, up to 100 paragraph ranges with a truncation flag, and whether the last paragraph is empty or a list item.

Create populated tables.

Debug table structures.

Create, rename, delete, or populate document tabs from markdown.

Read values from ranges as tab-separated rows, each labelled with its sheet row number.

Write, update, and clear range values, insert Drive and People smart chips, or copy another range into place.

Create spreadsheets.

List available spreadsheets.

Read spreadsheet metadata and sheet info.

Apply styling, number/text formats, cell merges, borders, and dropdowns to ranges.

List structured tables with IDs, names, ranges, and columns.

Add sheets to existing spreadsheets, or duplicate a tab within or into another spreadsheet.

Manage an existing tab: rename, delete, hide, unhide, reorder, or style it. Use create_sheet to add a tab and resize_sheet_dimensions for row and column changes.

Append rows to a structured table, auto-extending the table range.

Move rows between sheets within a spreadsheet.

Resize, hide, freeze, insert, or delete rows and columns.

List comments on spreadsheets.

Create, reply to, resolve, update, and delete spreadsheet comments. Pass cell (A1 notation) on create to anchor a comment to a specific cell. Update and delete apply to top-level comments you wrote.

Add, update, and remove conditional formatting rules.

List, create, update, and delete named ranges.

Create presentations.

Retrieve presentation details.

Apply multiple slide updates in one request.

Retrieve a specific page/slide.

Render a slide or notes page to an image and return it as image content.

List all presentation comments, including each comment's anchor so it can be matched to the slide element it refers to.

Create, reply to, resolve, update, and delete presentation comments. Update and delete apply to top-level comments you wrote.

Create forms.

Retrieve form metadata and URLs.

List responses with pagination.

Configure whether a form is published and accepting responses.

Retrieve an individual response.

Apply batch updates to form structure/settings.

List tasks with filters.

Retrieve task details.

Consolidated create, update, delete, and move task operations.

List task lists.

Retrieve task-list details.

Consolidated create/update/delete/clear-completed for task lists.

Search by name, email, and phone.

Retrieve detailed contact profiles.

List contacts with pagination.

Create, update, or delete contacts.

List contact groups.

Retrieve group details and members.

Batch create, update, or delete contacts.

Create, update, delete groups, and manage membership.

Retrieve messages in a space.

Send messages to spaces, or edit a message already sent.

Search chat history.

Add emoji reactions to messages.

List spaces and DMs, or the members of one space. Direct messages and unnamed group chats are labelled after their members.

Download chat attachments.

Run custom search queries with optional site restrictions.

Retrieve search engine configuration metadata.

List script projects, or read one project's structure or a single source file.

Create standalone or bound script projects, or permanently delete one.

Create or update script files.

Execute deployed script functions with args.

Generate Apps Script code for time-driven and event triggers.

List deployments for a project.

Create, update, or delete deployments. Updates keep the deployment's existing settings and change only the fields you pass.

List saved versions of a script project, or retrieve one.

Create an immutable version from the current project state.

Inspect recent executions, or aggregate execution metrics for a script.

List or delete your installable triggers on a project, via a small helper file run through the Execution API.