Smartsheet MCP Server
Enables AI agents to read and write Smartsheet data, with support for remote HTTP deployment to Microsoft 365 Copilot and Copilot Studio.
README
Smartsheet MCP Server
A production-ready Model Context Protocol (MCP) server that gives AI agents full read/write access to Smartsheet — equivalent to the native Claude MCP, but deployable as a remote HTTP endpoint for Microsoft 365 Copilot declarative agents and Copilot Studio.
Tools Exposed
| Tool | Description | Read-only |
|---|---|---|
smartsheet_list_sheets |
List all accessible sheets | ✅ |
smartsheet_get_sheet |
Read sheet data (columns + rows + cells) with filters & pagination | ✅ |
smartsheet_get_columns |
Get column definitions and IDs for a sheet | ✅ |
smartsheet_add_rows |
Add rows with cell values and hierarchy placement | ❌ |
smartsheet_update_rows |
Update existing row cells, lock/unlock rows | ❌ |
smartsheet_delete_rows |
Permanently delete rows | ❌ |
smartsheet_add_columns |
Add new columns with type/options | ❌ |
smartsheet_get_report |
Read paginated report data | ✅ |
smartsheet_list_reports |
List all accessible reports | ✅ |
smartsheet_list_workspaces |
List workspaces | ✅ |
smartsheet_browse_workspace |
Browse all contents of a workspace | ✅ |
smartsheet_browse_folder |
Browse folder contents | ✅ |
smartsheet_list_dashboards |
List all dashboards (Sights) | ✅ |
smartsheet_list_sheet_discussions |
Get all discussions on a sheet | ✅ |
smartsheet_list_row_discussions |
Get discussions on a specific row | ✅ |
smartsheet_create_sheet_discussion |
Start a new sheet-level discussion | ❌ |
smartsheet_create_row_discussion |
Start a new row-level discussion | ❌ |
smartsheet_add_comment |
Reply to an existing discussion | ❌ |
smartsheet_delete_comment |
Delete a comment (own comments only) | ❌ |
smartsheet_list_attachments |
List all attachments on a sheet | ✅ |
smartsheet_list_row_attachments |
List attachments on a row | ✅ |
smartsheet_attach_url_to_row |
Attach a URL link to a row | ❌ |
smartsheet_delete_attachment |
Permanently delete an attachment | ❌ |
smartsheet_search |
Full-text search across all Smartsheet assets | ✅ |
smartsheet_search_sheet |
Full-text search within one sheet | ✅ |
smartsheet_get_cell_history |
Audit trail — all historic values of a cell | ✅ |
smartsheet_get_sheet_version |
Check if a sheet was modified (lightweight) | ✅ |
Prerequisites
- Node.js 18+
- A Smartsheet API token — generate one at: Account → Apps & Integrations → API Access
- For M365 Copilot deployment: an Azure App Service or Azure Container Apps to host the HTTP server
Local Development (stdio — Claude Desktop or Copilot Studio local)
npm install
npm run build
# Set your token
export SMARTSHEET_API_TOKEN=your_token_here
# Run in stdio mode (default)
npm start
Claude Desktop config (claude_desktop_config.json)
{
"mcpServers": {
"smartsheet": {
"command": "node",
"args": ["/absolute/path/to/smartsheet-mcp-server/dist/index.js"],
"env": {
"SMARTSHEET_API_TOKEN": "your_token_here"
}
}
}
}
Remote HTTP Deployment (for M365 Copilot)
# Run in HTTP mode
TRANSPORT=http SMARTSHEET_API_TOKEN=your_token_here npm start
# Server listens at http://localhost:3000/mcp
# Health check: GET http://localhost:3000/health
Per-request Token (multi-user / OAuth scenarios)
Each request can pass its own token via header, overriding the server default:
X-Smartsheet-Token: user_specific_token
This supports SSO scenarios where each M365 user authenticates to Smartsheet with their own credentials.
Deploying to Azure App Service
# Build the project
npm run build
# Create a zip of the deployable files
zip -r deploy.zip dist/ package.json package-lock.json
# Deploy via Azure CLI
az webapp deploy --resource-group <rg> --name <app-name> --src-path deploy.zip
# Set environment variables in Azure
az webapp config appsettings set \
--resource-group <rg> \
--name <app-name> \
--settings SMARTSHEET_API_TOKEN=your_token TRANSPORT=http
Your MCP endpoint will be:
https://<app-name>.azurewebsites.net/mcp
Wiring into M365 Copilot (Declarative Agent)
Option A — Copilot Studio (Recommended for EPO/no-code)
- Open Copilot Studio → Create or edit an agent
- Go to Actions → Add an action → Model Context Protocol (MCP)
- Enter your server URL:
https://<app-name>.azurewebsites.net/mcp - Select the tools you want the agent to use
- Configure authentication (API key or OAuth)
- Publish the agent
ℹ️ MCP is now generally available in Copilot Studio as of mid-2025.
Option B — VS Code + Microsoft 365 Agents Toolkit (for IT/Dev)
- Install Microsoft 365 Agents Toolkit extension in VS Code
- Create a new Declarative Agent from the toolkit
- Choose Add Action → Start with an MCP server
- Enter your MCP endpoint URL
- The toolkit auto-generates the plugin manifest by reading your tool schemas
- Pick which tools to expose
- Configure OAuth (Smartsheet supports OAuth 2.0) or API key auth
- Deploy to your M365 tenant via Microsoft 365 Admin Center → Copilot → Agents
Option C — Direct Manifest (Advanced)
The Agents Toolkit auto-generates ai-plugin.json and openapi.json from your MCP server's tool list. These files go into the declarative agent's appPackage/ directory and are uploaded to the Teams App Catalog.
Authentication Options for M365
| Method | When to use |
|---|---|
API Key in header (X-Smartsheet-Token) |
Single shared service account token; simplest setup |
| OAuth 2.0 (per user) | Each user authenticates with their own Smartsheet account; required for per-user audit trails |
| Managed Identity + Key Vault | Best practice for production Azure deployments; store token in Key Vault, bind to App Service MSI |
Smartsheet OAuth App Setup (for per-user auth)
- Go to Smartsheet Developer Portal
- Create a new app → Set redirect URI to
https://teams.microsoft.com/api/platform/v1.0/oAuthRedirect - Copy the Client ID and Client Secret
- In Agents Toolkit: choose OAuth (static registration) and paste these values
Environment Variables
| Variable | Required | Description |
|---|---|---|
SMARTSHEET_API_TOKEN |
Yes (unless using per-request header) | Smartsheet API Bearer token |
TRANSPORT |
No (default: stdio) |
Set to http for remote/Copilot deployments |
PORT |
No (default: 3000) |
HTTP listen port |
Known M365 Copilot Limitations
- Only tools are supported — MCP resources and prompts are ignored by Copilot
- Nested object schemas with
minimum/maximum/defaulton nested properties can fail manifest validation in the Agents Toolkit — strip these if provisioning fails - Max 5 tools injected inline when ≤5 plugins are defined in a declarative agent manifest; above 5, the orchestrator selects dynamically
- Confirmation prompts: Read-only tools (annotated
readOnlyHint: true) don't require user confirmation; write tools do on first call
Project Structure
smartsheet-mcp-server/
├── src/
│ ├── index.ts # Entry point, transport selection
│ ├── types.ts # Smartsheet entity types
│ ├── constants.ts # API base URL, limits
│ ├── services/
│ │ └── smartsheet.ts # API client, auth, error handling
│ └── tools/
│ ├── sheets.ts # Sheet CRUD, rows, columns
│ ├── reports-workspaces.ts # Reports, workspaces, folders, dashboards
│ ├── discussions-attachments.ts # Comments, attachments
│ └── search.ts # Search, cell history, version check
├── dist/ # Compiled JS (after npm run build)
├── package.json
├── tsconfig.json
└── README.md
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.