OpenBSP MCP for Google Workspace
A hosted MCP server that allows AI assistants to access Google Calendar and Sheets via OAuth proxy, requiring no Google Cloud setup.
README
OpenBSP MCP for Google Workspace
A public Model Context Protocol (MCP) server that lets AI assistants access your Google Calendar and Sheets.
Live at: https://g.mcp.openbsp.dev
Why this exists
Most Google MCP servers require you to create your own Google Cloud project, set up OAuth credentials, and run a local server. This is tedious and error-prone.
This project is different. It's a hosted public server with proxy OAuth. You just:
- Click "Login with Google"
- Get an API key
- Add it to your MCP client
No Google Cloud setup. No local servers. No environment variables. Just connect and go.
What you can do
Once connected, ask your AI assistant to:
- List and search your calendar events
- Create, update, or delete events
- Read and write to Google Sheets
- Create new spreadsheets
How it works
┌──────────────┐ ┌──────────────────┐ ┌──────────────┐
│ AI Client │────▶│ OpenBSP MCP │────▶│ Google APIs │
│ (Claude) │◀────│ Server │◀────│ │
└──────────────┘ └──────────────────┘ └──────────────┘
- You select which Google products to connect (Calendar, Sheets)
- You authenticate with Google and grant permissions
- You receive an API key to configure your MCP client
- Your AI assistant can now access your Google data
Privacy & Security
What we store
| Data | Purpose | Protection |
|---|---|---|
| API Key | Authenticate your MCP client | Stored as-is (random 256-bit value) |
| OAuth Tokens | Access Google APIs on your behalf | AES-256-GCM encrypted |
| Scopes | Remember which products you authorized | Not encrypted |
OAuth tokens are encrypted at rest using AES-256-GCM. The encryption key is stored separately from the database, so a database breach alone cannot expose your Google tokens.
What we DON'T store
- Your Google password
- Your email content
- Your calendar events or spreadsheet data
- Any personal information
Data retention
- API keys: Stored until you revoke them (or until expiration if set)
- OAuth tokens: Stored alongside your API key
- Temporary auth states: Deleted after 10 minutes
How to revoke access
Option 1: Delete your API key
curl -X DELETE https://g.mcp.openbsp.dev/key/YOUR_API_KEY
Option 2: Revoke from Google
- Go to Google Account Security
- Find "OpenBSP MCP" in the list
- Click "Remove Access"
Both options will immediately invalidate your API key.
Setup for MCP Clients
After authenticating, add this to your MCP client configuration:
{
"mcpServers": {
"google-mcp": {
"url": "https://g.mcp.openbsp.dev/mcp",
"headers": {
"Authorization": "Bearer YOUR_API_KEY"
}
}
}
}
Programmatic integration
MCP clients can automate the setup by redirecting users to:
https://g.mcp.openbsp.dev/auth/google?products=calendar,sheets&callback=YOUR_CALLBACK_URL
After authentication, the user is redirected to your callback with credentials in the URL fragment:
YOUR_CALLBACK_URL#api_key=gmc_xxx&url=https://g.mcp.openbsp.dev/mcp&email=user@gmail.com&files=Budget.xlsx,Sales.xlsx
| Parameter | Description |
|---|---|
api_key |
The generated API key |
url |
The MCP endpoint URL |
email |
The Google account email that authorized access |
files |
Comma-separated file names (only when Sheets authorized and files selected) |
Available Tools
Calendar
Scopes requested:
calendar.events- read and write eventscalendar.calendarlist.readonly- list your calendars (read-only)
| Tool | Description |
|---|---|
list_calendars |
List all your calendars |
list_events |
Get events with optional date filters |
check_availability |
Check busy/free times (no event details exposed) |
create_event |
Create a new calendar event |
update_event |
Update an existing event |
delete_event |
Delete an event |
Sheets
Scope requested:
drive.file- access only files you select or files created by the app
| Tool | Description |
|---|---|
list_authorized_files |
List files the user has authorized for access |
get_spreadsheet |
Get spreadsheet metadata, sheets, and tables |
get_sheet_schema |
Get column names and inferred types |
describe_sheet |
Get statistical summary of all columns |
search_rows |
Search for rows matching filter criteria |
semantic_search |
AI-powered semantic search over rows (finds by meaning, not exact match) |
read_sheet |
Read data from a range (e.g., "A1:D10") |
write_sheet |
Write data to a range |
append_rows |
Append rows (supports Tables with tableId for footer-aware append) |
create_spreadsheet |
Create a new spreadsheet |
The drive.file scope only grants access to files the user explicitly selects during authentication (via Google Drive Picker) or files created by the app. Use list_authorized_files to discover which spreadsheets are available.
Tables support: The get_spreadsheet tool returns table metadata including tableId. Pass this to append_rows for proper table-aware appending that respects footers and auto-expands the table.
Semantic search: The semantic_search tool uses AI embeddings to find rows by meaning. Powered by TAG.
Self-Hosting
Want to run your own instance? See DEPLOY.md for instructions.
FAQ
Is this official? No, this is an independent open-source project. It is not affiliated with Google or Anthropic.
Is it safe? We only request the minimum permissions needed. Your data flows directly between Google and your AI client through our server - we don't store or log it.
Can I see the code? Yes! This project is open source. Review the code, run your own instance, or contribute improvements.
What if I lose my API key? Generate a new one by authenticating again. Old keys remain valid unless you revoke them.
Does it cost anything? This public instance is free. If you self-host, you'll pay for Cloudflare Workers (generous free tier available).
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.