drive-update-mcp
Enables Claude to update existing Google Drive files in place via a service account, preserving file ID, link, and revision history. Supports finding, creating, listing revisions, and updating files through a Streamable HTTP MCP server.
README
drive-update-mcp
A small remote MCP server that gives Claude one capability the built-in Google Drive connector is missing: updating an existing Drive file in place.
When Claude calls update_drive_file, the file keeps the same file ID, the same
shareable link, and Drive retains the previous version in its revision history. No
duplicates, no broken links.
How it works
Claude (claude.ai / app) ──► this MCP server (on Render) ──► Google Drive API
files.update() ← overwrites in place
- Claude connects to the server from Anthropic's cloud using the Streamable HTTP MCP transport.
- The server authenticates to Google Drive with a service account — no interactive Google login to manage. You grant it access to a file by sharing that file (or its folder) with the service-account's email address.
Tools exposed
| Tool | What it does |
|---|---|
update_drive_file |
Overwrite an existing file's contents by fileId (in place). |
find_files |
Search by name to resolve a file's fileId. |
create_drive_file |
Create a new file from base64 content (optionally in a folder). |
list_revisions |
Show a file's revision history. |
Prerequisites
- A Google account with Google Drive.
- A Google Cloud project (free).
- A Render account (free tier is fine) — or any Node host that gives you an HTTPS URL.
- A Claude Pro, Max, Team, or Enterprise plan (custom connectors are not on Free).
Setup
Part A — Create the Google service account
- Go to the Google Cloud Console → create or pick a project.
- APIs & Services → Library → enable the Google Drive API.
- IAM & Admin → Service Accounts → Create service account (any name, e.g.
drive-updater). - Open the service account → Keys → Add key → Create new key → JSON. A
.jsonfile downloads. - Note the service account's email (looks like
drive-updater@your-project.iam.gserviceaccount.com).
Part B — Give the service account access to your file
The service account can only touch files that are shared with it.
- In Google Drive, share the target file (or, easier, the folder that holds it) with the service-account email from step A5, with Editor permission.
- Do this once per file/folder you want Claude to be able to update.
Tip: keep all the docs Claude manages in one folder and share just that folder.
Part C — Deploy to Render
- Push this folder to a GitHub repo (or use Render's "deploy from local" / blueprint).
- In Render: New → Web Service → point it at the repo.
render.yamlis included, so Render will pick up the build/start commands automatically. Otherwise set:- Build command:
npm install - Start command:
npm start
- Build command:
- Add two Environment Variables:
GOOGLE_SERVICE_ACCOUNT_JSON→ paste the entire contents of the JSON key file from Part A4.MCP_AUTH_TOKEN→ a long random string you invent (this protects the endpoint).
- Deploy. When it's live, confirm
https://YOUR-APP.onrender.com/healthreturns{"ok":true,...}.
Part D — Connect it to Claude
- In Claude: Settings → Connectors → "+" → Add custom connector (on Pro/Max you may need to enable connector/developer features first).
- Name:
Drive Update - URL:
https://YOUR-APP.onrender.com/mcp?token=YOUR_MCP_AUTH_TOKEN(the?token=...is how the server authorises the call — use the same value asMCP_AUTH_TOKEN). - Click Add. Leave the OAuth fields blank.
- In a conversation, enable it via the "+" → Connectors toggle.
Using it with Claude
Once connected, the typical flow is:
- You ask Claude to make changes to a document it produces.
- Claude builds the updated file, then calls
find_files(to get thefileIdfrom the name) or you paste the Drive link, andupdate_drive_fileto overwrite it in place.
For the very first upload, the file has to exist in Drive (so it has an ID to update).
Either upload it once yourself, or have Claude call create_drive_file once; after
that, every change is an in-place update_drive_file.
Note on size: the new content is passed to the tool base64-encoded, so very large files mean large tool calls. For typical Office documents (well under a few MB) this is fine.
Security notes
- The
MCP_AUTH_TOKENin the URL is what stops anyone else from driving your server. Treat that URL like a password; rotate the token if it leaks. - The service account can reach only the files/folders you explicitly shared with it. It cannot see the rest of your Drive.
- For a hardened setup you can replace the token gate with a full OAuth flow (Claude supports OAuth + Dynamic Client Registration); that's a later upgrade, not needed to start.
Local development
npm install
# put GOOGLE_SERVICE_ACCOUNT_JSON and MCP_AUTH_TOKEN in your shell env (or a .env loader)
npm start
# health check:
curl localhost:3000/health
drive-update-mcp-
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.
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.
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.
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.