MCP · Data
Google Sheets
Claude reads and writes in your Google Sheets. Perfect to pipe data produced by another Skill into a tracker, or to analyze a CSV without opening it.
Use cases
- ›Update a weekly metrics tracker
- ›Analyze an export without Excel
- ›Build a dashboard across multiple sources
Installation
1.Create a Google Cloud project, enable the Sheets API, download credentials.json as OAuth Desktop. Place the file at ~/.config/workspace-mcp/credentials.json. The first call will open a browser to authorize access.
2.One line, one terminal
$ claude mcp add workspace -- npx -y workspace-mcpNeed to install Claude Code first? See the Claude Code page · Official docs
Concrete examples
›
"Dans le sheet 'Revenue Tracker 2026' onglet Novembre, ajoute une ligne avec la date d'aujourd'hui, 24 350€ de MRR et un commentaire '+3 clients enterprise cette semaine'."Row-by-row write with typed values. Claude calls sheets_append_row after locating the correct tab.
Sample output
→ Finding spreadsheet "Revenue Tracker 2026"... found → Finding tab "November"... found (next empty row: 23) → Appending row [2026-04-16, 24350, "+3 enterprise clients this week"] ✓ Row 23 added. Delta vs last week: +$1,850 (+8.2%). Monthly target $25,000: 97.4% achieved.
›
"Analyse l'onglet 'Churn' du sheet Clients : quel est le mois avec le plus de churn et pourquoi selon les commentaires ?"Read + aggregation. Claude sheets_read, groups by month, then reads the comments column to surface reasons.
Sample output
→ Reading Clients.Churn (127 rows) Churn by month (clients lost) Jan 4 Feb 6 Mar 5 Apr 12 May 3 Jun 4 Jul 2 Aug 5 Sep 7 Peak: April (12 lost clients, 3x the median). Reasons listed (Comments column) - 7/12: "too expensive after price bump" (announced March 15) - 3/12: "missing feature X vs competitor" - 2/12: no reason given Conclusion: your March price hike cost ~7 clients. If you consider another increase, plan an explicit grandfathering clause.