Google Workspace

Built by Runlayer. Runlayer builds and maintains this connector. It runs on Runlayer Deploy, which must be enabled for your workspace.

Connect your AI assistant to Google Workspace services. All Google servers share the same Google OAuth setup process.

Server	Description
Google Calendar	Manage events, calendars, and schedules
Gmail	Read, search, send emails, and manage labels
Google Drive	Browse, search, and manage files and folders
Google Docs	Create, read, and edit documents
Google Sheets	Create and manage spreadsheets
Google Slides	Create and edit presentations
Google Forms	Create, edit, and manage forms; access forms and their responses

Available Tools

Google Calendar

Tool	Description
`list_calendars`	List all accessible calendars
`list_events`	List events with date filtering
`get_event`	Get event details
`create_event`	Create an event with attendees (required or optional), visibility, recurrence, and an optional Google Meet link
`quick_add`	Create event from natural language
`update_event`	Modify an event, including attendees and visibility, or add a Google Meet link
`respond_to_event`	Respond to an invitation (accept, decline, or tentative)
`delete_event`	Remove an event
`move_event`	Move event to different calendar
`create_calendar`	Create a new calendar
`update_calendar`	Modify calendar settings
`delete_calendar`	Remove a calendar
`get_free_busy`	Check availability

Set add_google_meet: true on create_event or update_event to add Google Meet. The response returns the join link when ready. If Meet creation is still pending or fails, the event remains saved and the response includes a warning; use get_event to check a pending link, or retry a failed Meet request with update_event on the same event. Existing Meet links are preserved. Omitting the flag or passing false leaves conference details unchanged. This option does not replace other conference providers. The calendar must support Google Meet.

Gmail

Tool	Description
`search`	Search emails using Gmail operators
`list_messages`	List recent messages
`fetch_message`	Get full email content (includes attachment metadata)
`fetch_attachment`	Download a message attachment as text or base64
`send_email`	Send a new email (supports attachments)
`reply_to_message`	Reply to an email thread (supports attachments)
`create_draft`	Create a draft (supports attachments)
`send_draft`	Send an existing draft
`list_drafts`	List all drafts
`list_labels`	List all labels
`add_labels`	Add labels to messages
`remove_labels`	Remove labels from messages
`create_label`	Create a new label
`mark_as_read`	Mark messages as read
`mark_as_unread`	Mark messages as unread
`archive_messages`	Archive messages

Google Drive

Tool	Description
`search`	Search files by name or content; name matches ranked first
`list_folder`	List folder contents
`get_metadata`	Get file/folder metadata
`update_metadata`	Update a file/folder’s description and custom properties
`fetch_file`	Read file content, including extracted PDF text by default (1 MB cap; set `allow_large_file` to read up to 10 MB)
`create_file`	Create a new file
`create_folder`	Create a new folder
`create_doc_from_markdown`	Create Google Doc from Markdown
`update_file`	Update file content
`update_doc_from_markdown`	Update Google Doc from Markdown
`rename_file`	Rename a file or folder
`copy_file`	Copy a file
`move_file`	Move a file
`delete_file`	Move file to trash
`share_file`	Share a file
`list_permissions`	List sharing permissions

For PDFs, fetch_file defaults to format: "auto" and returns readable text, capped at 100,000 characters including any truncation marker. The response reports truncated and extraction details: status (text_extracted, partial, or ocr_required), page counts, pages without text, and a warning. Parsing has a 30-second deadline. Scans and images are not read because OCR is disabled; blank pages also count as pages without text. Layout, tables, and reading order may not be preserved. Use format: "raw" only when you need exact PDF bytes. It returns base64, roughly 1.4 times the file size in characters, all of which consumes model context. Both formats retain the 1 MB default download limit and the 10 MB ceiling when allow_large_file: true; the 100,000-character output cap applies to extracted text only. Here, 1 MB means 1,048,576 bytes. Existing Google Drive connectors need to be redeployed to receive PDF text extraction.

Google Docs

Tool	Description
`search`	Search for documents
`fetch`	Get text across tabs, with optional native structure and formatting via `include_structure`
`create_file`	Create a new document from markdown or HTML
`copy_document`	Copy a native template with styles, tables, lists, tabs, headers and footers preserved
`replace_file_content`	Overwrite a document’s entire content
`insert_text`	Insert text at position
`replace_text`	Find and replace text
`delete_text`	Find and delete text
`delete_between_text`	Delete a whole block (a heading and everything under it, tables included) by naming the text at each end
`apply_doc_updates`	Apply raw Google Docs API requests in one batch
`format_text`	Find text and apply character formatting (bold, color, font, etc.)
`format_paragraphs`	Find text and apply paragraph styling (heading, alignment, indent, line & paragraph spacing, bullet and numbered lists)
`link_text`	Find text and turn it into a hyperlink
`get_table`	Read a table’s row/column counts, index ranges, and cell text without changing the document
`edit_table`	Edit a table (insert/delete rows & columns, set cell text, overwrite a whole row, write many cells at once, column width, cell styling including padding, delete the table), located by index or text
`insert_break`	Insert a page or section break
`add_comment`	Anchor a native comment to a UTF-16 text range, optionally targeting a tab or header/footer segment
`list_comments`	Read native comment threads, replies, anchor ranges and quoted text
`resolve_comment`	Resolve a comment
`list_tabs`	List document tabs
`fetch_tab`	Get content of a tab
`fetch_all_tabs`	Get all tab contents

For template-based generation, use copy_document, fill placeholders with replace_text or apply_doc_updates, then use fetch with include_structure: true to inspect the native formatting. Styles may inherit from named styles and defaults; structured data does not verify rendered pagination. Avoid replacing the whole document body when preserving template formatting. Native comments use the Google Docs API to attach to text. Pass the native UTF-16 indices from fetch.include_structure as docStartIndex and docEndIndex to add_comment; legacy startIndex and endIndex inputs are rejected to prevent misplaced anchors. Comment creation may notify collaborators. After an ambiguous copy or comment-write failure, read back the result before retrying to avoid duplicates. Comment pagination tokens from the previous Drive-based implementation must be discarded. The documented Google Docs API does not expose native checklist checked/unchecked state. Existing Google Docs connectors need to be redeployed to receive these capabilities.

Google Sheets

Tool	Description
`search`	Search for spreadsheets
`list`	List spreadsheets
`fetch`	Get spreadsheet data (values or underlying formulas)
`create`	Create a new spreadsheet
`update`	Update cell values
`update_ranges`	Update multiple ranges in a single call
`append`	Append rows
`insert_rows`	Insert blank rows mid-sheet, shifting existing rows down
`delete_rows`	Delete rows, shifting rows below up
`clear`	Clear values in a range, keeping rows and formatting
`add_sheet`	Add a new sheet tab
`get_metadata`	Get spreadsheet metadata
`set_row_visibility`	Hide or unhide rows
`group_rows`	Group rows (optionally collapsed)
`ungroup_rows`	Remove a row group
`autofill`	Extend a formula/series across a range
`format_cells`	Apply cell formatting (number format, font, color, alignment, wrap)
`set_column_width`	Set column width in pixels, or fit to contents
`set_row_height`	Set row height in pixels, or fit to contents
`create_filter`	Turn on filtering over a range so rows can be sorted and filtered
`clear_filter`	Remove a sheet’s filter

Google Slides

Tool	Description
`search`	Search presentations by name or content; name matches ranked first
`fetch`	Get presentation content
`list_slides`	List all slides
`get_slide`	Get slide details
`list_layouts`	List available layouts
`create`	Create a presentation
`create_from_markdown`	Create slides from Markdown
`create_from_template`	Copy a themed template deck (optionally clearing sample slides)
`duplicate`	Duplicate a presentation
`add_slide`	Add a new slide
`delete_slide`	Remove a slide
`duplicate_slide`	Duplicate a slide
`reorder_slide`	Change slide order
`add_text`	Add a text box
`update_text`	Modify text in a shape
`add_shape`	Add a shape
`add_image`	Insert an image
`add_table`	Insert a table
`add_line`	Add a line
`delete_element`	Remove an element
`style_text`	Apply text formatting
`set_paragraph_style`	Set paragraph alignment
`style_shape_advanced`	Apply shape styling
`set_slide_background`	Change background
`set_speaker_notes`	Add speaker notes

Google Forms

Tool	Description
`search`	Search for forms
`list`	List all forms
`get`	Get form structure and questions
`responses`	Get form responses
`create`	Create a new form (title only; add content with `batch_update`)
`batch_update`	Add, edit, reorder, or delete questions/items; update form info or settings

Setup

Create Google Cloud Project

Go to Google Cloud Console
Create a new project or select an existing one

Enable APIs

Go to APIs & Services > Library and enable the APIs you need:

Google Calendar API
Gmail API
Google Drive API
Google Docs API
Google Sheets API
Google Slides API
Google Forms API

Configure OAuth Consent Screen

Go to APIs & Services > OAuth consent screen
Select External or Internal
Fill in app name, support email, and developer contact
Add required scopes (see below)
Add test users if using External mode

Create OAuth Credentials

Go to APIs & Services > Credentials
Click Create Credentials > OAuth client ID
Select Web application
Add redirect URI: https://<your-tenant>.runlayer.com/oauth/callback
Copy the Client ID and Client Secret

Add Server in Runlayer

In Runlayer, go to Connectors, select a Google service, and click Add connector
Check Custom OAuth credentials
Enter your Client ID and Client Secret
Click Deploy Server

If you have the Manage security policies capability, Gmail and Google Calendar open Recommended security settings on the connector page after deploy. Review the per-tool overrides, then click Apply.

Authorize

Click Connect on the server page
Sign in with Google and grant permissions

OAuth Scopes

Add these scopes when configuring OAuth consent:

Service	Scopes
Calendar	`https://www.googleapis.com/auth/calendar.events`, `https://www.googleapis.com/auth/calendar.calendars`, `https://www.googleapis.com/auth/calendar.calendarlist.readonly`, `https://www.googleapis.com/auth/calendar.freebusy`
Gmail	`https://www.googleapis.com/auth/gmail.readonly`, `https://www.googleapis.com/auth/gmail.send`, `https://www.googleapis.com/auth/gmail.modify`
Drive	`https://www.googleapis.com/auth/drive`
Docs	`https://www.googleapis.com/auth/drive`, `https://www.googleapis.com/auth/documents`
Sheets	`https://www.googleapis.com/auth/drive.metadata.readonly`, `https://www.googleapis.com/auth/spreadsheets`
Slides	`https://www.googleapis.com/auth/drive`, `https://www.googleapis.com/auth/presentations`
Forms	`https://www.googleapis.com/auth/drive.metadata.readonly`, `https://www.googleapis.com/auth/forms.body`, `https://www.googleapis.com/auth/forms.responses.readonly`

All connectors additionally request openid and https://www.googleapis.com/auth/userinfo.email for user identity.

Changing scopes on an existing connector

When a connector’s scopes change (for example, Google Forms moving from read-only access to read and write, so it can create and edit forms), an already-deployed connector needs these steps in this order:

Add the new scope to your OAuth consent screen (External apps only)

If your app’s user type is External, add the new scope(s) in APIs & Services > OAuth consent screen so users can consent to them. Internal (Google Workspace) apps can request any scope without listing it here, so you can skip this step. You can remove any scope the new one supersedes (e.g. forms.body covers forms.body.readonly).

Redeploy the connector

Redeploy the server so it picks up the new image and requests the new scope set. Until you redeploy, nothing changes for your users. By default the connector requests its built-in scopes, so a redeploy is enough — unless you’ve overridden them (see the note below).

Users re-authorize

Each user re-authorizes through their own MCP client, the same way they first connected (signing in with Google and granting access). Existing authorizations keep their previously granted scopes until the user re-authorizes, so users pick up the new capability individually as they reconnect — it isn’t something an admin does on each user’s behalf.

The connector requests its built-in scopes by default, so most deployments only need the redeploy. The exception is if you set a GOOGLE_OAUTH_SCOPES environment variable on the connector, which overrides the built-in scopes. To check, open your connector in Runlayer and look at its environment variables: if GOOGLE_OAUTH_SCOPES is set, add the new scope to it (otherwise a redeploy won’t change what’s requested); if it isn’t set, there’s nothing to do.

Each user will need to re-authorize before they can use the new tools. Until they do, tools that need the added scope return permission errors for that user, while their existing capabilities keep working.

Adding a sensitive or restricted scope to a published/verified External app may require re-verification by Google (verification of a related scope does not carry over). Internal (Google Workspace) apps do not need re-verification.

Troubleshooting

Redirect URI mismatch

Verify the redirect URL in Google Cloud Console matches exactly: https://<your-tenant>.runlayer.com/oauth/callback (no trailing slash).

Access blocked during OAuth

Ensure OAuth consent screen is configured. Add yourself as a test user if using External mode.

API not enabled

Enable the specific API in Google Cloud Console > APIs & Services > Library.

Invalid grant

Tokens may have expired. Disconnect and reconnect the server in Runlayer.

Insufficient permissions

Re-authorize with updated scopes. Ensure the API is enabled in your project.

Enable

Control

Self-Hosted Deployment

Operations

Cookbook

Integrations

Connectors

Resources

Available Tools

Google Calendar

Gmail

Google Drive

Google Docs

Google Sheets

Google Slides

Google Forms

Setup

OAuth Scopes

Changing scopes on an existing connector

Troubleshooting

​Available Tools

​Google Calendar

​Gmail

​Google Drive

​Google Docs

​Google Sheets

​Google Slides

​Google Forms

​Setup

​OAuth Scopes

​Changing scopes on an existing connector

​Troubleshooting

Available Tools

Google Calendar

Gmail

Google Drive

Google Docs

Google Sheets

Google Slides

Google Forms

Setup

OAuth Scopes

Changing scopes on an existing connector

Troubleshooting