Google Docs MCP: Official, but Developer Preview Only
Google does have an official MCP server for Docs — but it’s not something you can just switch on. It ships under the Google Workspace Developer Preview Program, and using it means creating a Google Cloud project, enabling two APIs, and configuring an OAuth consent screen yourself. There’s no toggle in your regular Google account settings. So if you searched “Google Docs MCP” expecting a one-click connector, the honest answer is: it exists, it’s real, and it’s built for developers first.
Even once it’s running, an MCP server only ever hands Docs to an AI inside a conversation you start. It’s a doorway, not a worker. Nothing watches your documents for you, nothing fires when a doc changes, and nothing runs while the chat is closed. Here’s what the Google Docs MCP server actually does, what setting it up involves, where it stops even for developers who use it — and what to use instead if you just want Docs work to happen on its own.
Google Docs is one of many — the complete list of MCP servers covers every app that ships one, grouped by category.
What the Google Docs MCP server does
Model Context Protocol (MCP) is the open standard that lets an AI client — Claude, Google Antigravity, and others — talk to an outside app through a shared interface. Google’s own docs describe the Docs MCP server plainly: “Google Docs offers a remote Model Context Protocol (MCP) server that allows AI agents to securely interact with Google Docs data,” built to work with AI applications like Google Antigravity and Claude.
It exposes exactly two tools:
read_doc— retrieves a full JSON representation of a Google Doc by its document ID, including text and structural information. Under the hood it’s the same as the Docs API’sdocuments.get.update_doc— applies changes through a batch-update call (documents.batchUpdate), covering more than 40 operations: replacing text, changing styles, editing tables, adding comments and suggestions, and touching headers, footers, and document tabs.
That’s it. No third tool for search, no tool for creating a brand-new doc from scratch, no tool for watching a folder of docs. It reads one doc and writes to one doc, both by ID, both inside whatever chat session invoked it.
How to set up the Google Docs MCP server
This is the part that separates Docs from something like HubSpot’s remote MCP server, which any non-developer can point a client at in a few clicks. The Docs server requires actual Google Cloud setup:
- Create a Google Cloud project (or use an existing one).
- Enable two APIs — the Google Docs API (
docs.googleapis.com) and the Google Docs MCP API (docsmcp.googleapis.com) — viagcloudor the Cloud console. - Configure an OAuth consent screen — set an app name, support email, audience (internal or external, with test users added), and manually add the scopes the server needs:
drive.readonlyanddocuments.readonly. - Create an OAuth client ID and secret, then add the server’s endpoint (
https://docsmcp.googleapis.com/mcp) to your MCP client’s config — Google’s instructions currently cover Antigravity by name.
None of that is unreasonable for a developer. It is not something a person who just wants an AI to help with their Docs is going to do on a Tuesday afternoon. If you’d rather skip Google Cloud entirely, community-built servers like google_workspace_mcp bundle Docs alongside Sheets, Drive, and Gmail with a simpler OAuth flow — but those are third-party projects, not Google’s own, and you’re trusting their code with your account scopes.
Where the Google Docs MCP stops
Even fully configured, the same limits that apply to every MCP server apply here:
- It only works inside a chat you start. Close the window and nothing happens. The AI doesn’t watch your Docs; it waits for you to ask.
- No triggers. A comment left on a doc, a new version saved, a doc shared with you — none of these can start anything through MCP. There’s no “when this happens in Docs, do that.”
- It’s one app at a time. The Docs MCP server knows Docs. Getting the contents of a doc into Slack, a CRM, or an email means wiring up a separate MCP server for each and hoping your client can juggle them in one turn.
- You own the plumbing, the project, and the scopes. The Google Cloud project, the OAuth consent screen, the client secret, and the blast radius of
documents.readonlyanddrive.readonlyaccess are all on you to maintain.
So even for the developers who go through the setup, the Docs MCP server is a way to ask about a document and make one-off edits mid-conversation. It is not a way to make Docs work run — to have something happen on a schedule or in reaction to an event, across the other tools a document touches.
Running Google Docs work that doesn’t need a chat open
That “run on its own, across apps” gap is exactly where Carly fits — without any Google Cloud project, OAuth consent screen, or gcloud command. Carly connects to Google Docs natively, plus the ~260 other apps it supports and anything with a public API through your own key. The difference from MCP is the important part: Carly’s workflows are triggered and scheduled, so Docs work happens whether or not anyone has a chat window open.
A few things MCP can’t do but a Carly workflow can:
- When a new doc is created in a shared folder → summarize it and post the summary to the right Slack channel automatically.
- Every Friday → pull the week’s meeting-notes docs, draft a rollup, and drop it in an email for approval.
- When a doc is marked “final” → convert the key sections into a CRM note and notify the account owner.
The non-AI steps — the moving, matching, and routing between apps — are free and unlimited, the Zapier-style backbone of the workflow. The AI steps (drafting, summarizing, deciding) start at $35/month. You describe the outcome in plain language and Carly wires up the Docs connection and everything downstream.
If you’re a developer who wants to interrogate a document from inside Antigravity or Claude, Google’s own MCP server is the real, official tool for that — it’s just gated behind a Cloud project setup, not a switch in your Google account. If you want Docs to actually do things on a trigger or a schedule, across every app a document flows through, that’s the job MCP wasn’t built for, and it’s the one Carly was.
FAQ
Does Google Docs have an official MCP server?
Yes. Google publishes an official remote MCP server for Docs at docsmcp.googleapis.com, currently under the Google Workspace Developer Preview Program. It offers two tools, read_doc and update_doc, and is documented for use with clients like Google Antigravity and Claude.
Can I just turn on the Google Docs MCP server from my Google account? No. Unlike a simple account connection, it requires creating a Google Cloud project, enabling the Docs API and the Docs MCP API, and configuring an OAuth consent screen with specific scopes. It’s built for developers, not a one-click consumer feature.
Can the Google Docs MCP server trigger automations? No. MCP is request/response inside an AI chat — it has no triggers and nothing runs when the conversation is closed. For event- or schedule-driven Docs work across apps, you need a workflow tool like Carly rather than an MCP server.
Can I connect Google Docs to AI without coding or hosting a server? Yes. You don’t have to touch Google Cloud, OAuth consent screens, or MCP at all. Carly connects to Google Docs for you and lets you build the automation in plain language — describe what you want to happen and it wires up Docs and the other apps involved, with no server to host and no code to write.
Ready to automate your busywork?
Carly schedules, researches, and briefs you—so you can focus on what matters.
See what people say
"Before Carly, I relied on a Calendly link, but the whole process felt impersonal and not very professional. Carly changed that by handling all the back-and-forth, so I'm no longer stuck in endless email threads trying to line up schedules.
Now Carly reaches out to candidates, shares my real-time availability, lets them pick a slot, then sends a Zoom link and drops it straight into my calendar. She sends reminders to both of us before each call, which has significantly reduced no-shows and last-minute confusion.
On top of scheduling, Carly acts like a full executive assistant, sending me my schedule the night before so I can prepare for each call. It reminds me of the old x.ai assistant, but Carly is noticeably smarter, faster, and better suited to my healthcare recruitment business."
