Google Sheets MCP: connect Google Sheets to Claude and ChatGPT
Ask Claude or ChatGPT which cost lines in the 2027 budget are over plan, how many people have signed up for the October customer breakfast, or to add this week’s hours to the project time log. The AI reads the exact sheet and range in Google Sheets, row by row, and gives you the link.
With Deployed OS everyone connects their own Google account. The AI reads the spreadsheets that person can already open. With Edit access it appends rows at the end of a sheet or creates a new spreadsheet, after they confirm the values. No cell is ever overwritten.
3 read actions · 2 write actions · Works in Claude, ChatGPT and Copilot
What your team can do with Google Sheets in Claude and ChatGPT
Any tab, not just the first
A new row in one sentence
Count the sign-ups
A new sheet from the conversation
Ask it like this
“Which tabs are in the 2027 budget spreadsheet, and how many rows does each have?”
“Read Forecast!A1:H60 in the sales forecast and tell me which regions are below target.”
“Add a row to the project time log: 3 October, Fjällgren Bygg, workshop prep, 2.5 hours.”
“How many people have signed up for the October customer breakfast, and from which companies?”
“Create a spreadsheet called “Supplier comparison” with columns for supplier, lead time, minimum order and contact, and the three suppliers we just discussed.”
“In the stock list, which items have fewer than 20 left?”
What the Google Sheets MCP connector does today
- Reads
sheets_search_spreadsheetsFinds the Google Sheets you can open by words in their names. Returns name, owner, last change and link for each.
- Reads
sheets_get_spreadsheetReads a spreadsheet’s title and its sheets, with each sheet’s name and size, so the right range can be picked.
- Reads
sheets_read_rangeReads a range as the values you see, row by row, in A1 notation such as Budget!A1:D50, or the first sheet when no range is given. At most 1,000 rows.
- Writes
sheets_append_rowsAppends rows at the end of a sheet in a spreadsheet you may edit, once you have confirmed the values. Values are stored as typed, never as formulas, nothing is overwritten and nobody is notified.
- Writes
sheets_create_spreadsheetCreates a spreadsheet in your own Drive with a title and, if you want, a sheet name and rows, once you have confirmed them. It is shared with nobody.
Need it to do more? We extend connectors for the companies that use them.
Who sees what
- Reads, plus new rows
- The connector finds spreadsheets by name, lists their tabs and reads any range as the values you see. With Edit it can append rows or create a spreadsheet. It cannot overwrite, clear or format cells, write formulas, or share or delete anything.
- Each person’s own rights
- Everyone connects their own Google account, so the AI reads what that person can open and appends only where they may edit. The salary sheet shared with HR stays with HR.
- Writing only with Edit access
- Admins choose which groups get Google Sheets, and only groups with Edit access can append rows or create spreadsheets. Each write is done as the person asking, after they confirm the values, and nobody is notified.
- Every call logged
- Who read which range, who added which rows, when. Deployed OS shows reads and writes per connector and per person, the log can be exported, and access is revoked centrally the day someone leaves.
Set up the Google Sheets connector in three steps
Google Sheets’s own MCP server or Deployed OS?
Google runs its own remote MCP server for Sheets, one of its Workspace MCP servers, available through the Google Workspace Developer Preview Program. A developer turns it on in a Google Cloud project and connects an AI client, which can then read values and update cells and formulas. For one person, or a developer building an agent, that is a good fit.
Deployed OS is built for the whole company: one connection for Google Sheets and the rest of your tools, each person’s own login, admins choosing which groups may read and which may write, and a log of every call. The same connection works in Claude, ChatGPT and Copilot, next to your company’s skills and guardrails.
Google Sheets MCP: common questions
Yes. Google has a remote MCP server for Sheets in its Workspace Developer Preview Program, which a developer sets up in a Google Cloud project. Deployed OS gives the whole company one connection instead, with each person’s own login, access per group and a log of every call.
No. With Edit access it can append rows at the end of a sheet or create a new spreadsheet in your own Drive, when you ask and confirm the values. It never overwrites, clears or formats cells, and values are stored as typed, so a cell starting with = stays text rather than becoming a formula. If you want it to do more, for example update a cell, we can extend it for you.
Your admin sets up Google Sheets in Deployed OS once. You add your company’s Deployed OS connection in Claude, connect your own Google work account and start asking. The same connection works in ChatGPT and Copilot.
The Google Drive connector reads a spreadsheet as the CSV of its first tab, read-only. The Google Sheets connector lists every tab, reads any range as the values you see, up to 1,000 rows at a time, and with Edit access adds rows or creates spreadsheets. You can turn on both.
Only the ones they can already open in Google Sheets. The AI uses each person’s own Google account, so your sharing settings apply as they are, and a new spreadsheet is shared with nobody until its owner shares it.
Deployed is an EU company and fully GDPR compliant, and you choose where your data is stored, down to the city. On-premises is possible too. Spreadsheets are fetched from Google Sheets when someone asks for them.
It depends on the size of your team. Get in touch and we will walk you through it.
See it on your own tools.
Tell us which systems your company runs on. We’ll show you what your team can ask, and what IT gets to control.
