Talk to your Iterable data using natural language. Ask questions or give instructions like "How many campaigns did we send last week?", "Show me my most recent templates", or "Build me a beautiful email template that does the following..." and get instant answers without writing code or navigating dashboards.
Supported AI clients:
Model Context Protocol (MCP) is an open standard that lets AI assistants securely connect to external tools and APIs. This server acts as a bridge between your AI client and Iterable, translating natural language requests into safe API calls.
Prerequisites:
- Node.js >= 20 (v22 LTS recommended)
- An Iterable API key
npx @iterable/mcp setupTo always use the latest version (auto-update on each restart), add --auto-update:
npx @iterable/mcp setup --auto-updateBy default, the setup wizard configures the server in a safe, read‑only mode (no PII tools, no writes, no sends). To selectively enable elevated capabilities during setup, pass --advanced:
# Example: enable advanced setup to configure permissions
npx @iterable/mcp setup --advancedWhat you'll choose (optional):
- User PII (
ITERABLE_USER_PII): access user profile data, including email addresses, phone numbers, and custom data fields. - Writes (
ITERABLE_ENABLE_WRITES): create, update, and delete resources such as templates, lists, catalogs, campaigns, snippets, and user profiles. - Sends (
ITERABLE_ENABLE_SENDS): send messages (email, SMS, push, in-app, WhatsApp), trigger campaigns and journeys, schedule and abort campaigns, and track events. Requires writes to be enabled. Note: creating a blast campaign will schedule it for delivery, matching the behavior of the underlying Iterable API; there is no way to create a draft campaign. If you only need to draft content, you can do so with sends disabled by working with templates instead.
IMPORTANT: Enabling writes and sends allows the AI agent to take real, potentially irreversible actions against your Iterable project, including sending messages to real users and deleting data. If you do not have the technical knowledge to properly review the agent's tool calls before they are executed, you should avoid enabling these flags, especially in production environments. It is entirely your choice to accept this risk. If you enable these capabilities, it is your responsibility to carefully review each action before allowing the agent to proceed.
Note that permission settings are saved per key (see key management section below), allowing you to enable different permissions for different projects, e.g. only enable writes and sends for a sandbox project and disable them in production.
See TOOLS.md for all available tools with descriptions. All tools map directly to Iterable API endpoints.
Try these prompts:
- "Get details on campaign 12345"
- "What email templates are available?"
- "Show me all my product catalogs"
- "What journeys are currently active?"
- "Show me events for user@example.com from the last 30 days"
- "Create a campaign called 'Holiday Sale' using template 456"
- "Export all user data from January 2024"
- "List users in my 'VIP Customers' list"
Key Storage:
API keys are stored in the ~/.iterable-mcp/keys.json file and managed via the npx @iterable/mcp keys commands. On macOS the actual API key values are stored in the system Keychain. On Windows, API key values are encrypted using the Windows Data Protection API (DPAPI). On Linux, the API key values are stored directly in the file with restricted permissions (0o600).
Each key is tied to its API endpoint (US, EU, or custom) and to its permissions (view PII, write operations, send messages).
How Key Selection Works:
- You can store multiple API keys with different names (e.g., "production", "staging", "dev")
- Only ONE key is marked as active at a time
- The MCP server automatically uses whichever key is currently active
- Your first key is automatically set as active
- Switch between keys using the
activatecommand
# List stored keys (shows which one is active with ● ACTIVE badge)
npx @iterable/mcp keys list
# Add a new key (interactive: prompts for name, region/endpoint, and API key)
# Your first key becomes active automatically
npx @iterable/mcp keys add
# Switch to a different key by name or ID (also switches endpoint)
npx @iterable/mcp keys activate production
npx @iterable/mcp keys activate staging
# Delete a key by ID (requires ID for safety)
# Note: Cannot delete the currently active key - activate another first
npx @iterable/mcp keys delete <key-id>
# To update a key: delete the old one and add a new one with the same nameIf you'd rather not use npx, you can install globally. This lets you use iterable-mcp as a shorthand for npx @iterable/mcp.
pnpm add -g @iterable/mcp
iterable-mcp setupNote: The setup command automatically configures the correct command path.
The setup --claude-code command automatically configures Claude Code by running claude mcp add for you.
Alternatively, you can configure it manually:
# Add your API key first (see API Key Management section above)
npx @iterable/mcp keys add
# Then configure Claude Code
claude mcp add iterable -- npx -y @iterable/mcp
# Verify it was added
claude mcp list
# Optional: Configure privacy settings
claude mcp add iterable \
--env ITERABLE_USER_PII=false \
--env ITERABLE_ENABLE_WRITES=false \
-- npx -y @iterable/mcpIf you have already configured Claude Desktop successfully, you can run this command to copy your MCP server settings to Claude Code:
claude mcp add-from-claude-desktopNote: All of the above claude mcp commands will save the settings to ~/.claude.json, which makes the configured MCP servers available across all projects. For shared, project-specific MCP configs, create a .mcp.json file in the root of your project.
For more information, see the Claude Code MCP documentation.
The above commands will automatically configure your AI tool to use the MCP server by editing the appropriate configuration file, but you can also manually edit the appropriate configuration file:
- Claude Desktop:
~/Library/Application Support/Claude/claude_desktop_config.json - Cursor:
~/.cursor/mcp.json - Windsurf:
~/.codeium/windsurf/mcp_config.json - Antigravity:
~/.gemini/antigravity/mcp_config.json - Gemini CLI:
~/.gemini/settings.json
All five use the same configuration format:
Recommended: Using key manager:
# First, add your API key (interactive prompts)
npx @iterable/mcp keys addThen edit your config file:
{
"mcpServers": {
"iterable": {
"command": "npx",
"args": ["-y", "@iterable/mcp"]
}
}
}No env section needed - API key and base URL are loaded automatically.
Alternative: Environment variables:
{
"mcpServers": {
"iterable": {
"command": "npx",
"args": ["-y", "@iterable/mcp"],
"env": {
"ITERABLE_API_KEY": "your-iterable-api-key-here",
"ITERABLE_BASE_URL": "https://api.iterable.com"
}
}
}
}If you need to use a custom path to node or npx (for example, if they are not in your PATH), you can set environment variables before running the setup command:
# Set custom paths
export ITERABLE_MCP_NODE_PATH="/path/to/custom/node"
export ITERABLE_MCP_NPX_PATH="/path/to/custom/npx"
# Then run setup
npx @iterable/mcp setup --cursorAlternatively, you can manually edit your configuration file (after adding your key):
{
"mcpServers": {
"iterable": {
"command": "/custom/path/to/npx",
"args": ["-y", "@iterable/mcp"]
}
}
}No env section is needed if using the key manager.
The setup wizard and key manager handle most of these automatically. Setting environment variables directly is useful for CI/CD pipelines, Docker containers, or other non-interactive environments where the key manager isn't available. When both are present, key manager settings take precedence over environment variables.
| Variable | Required | Description |
|---|---|---|
ITERABLE_API_KEY |
No* | Your Iterable API key (*Optional if using key manager, otherwise required) |
ITERABLE_BASE_URL |
No** | Base URL for the Iterable API (**Not needed when using key manager - URL is stored with each key) |
ITERABLE_DEBUG |
No | Set to true for API request logging |
LOG_LEVEL |
No | Set to debug for troubleshooting |
ITERABLE_USER_PII |
No | Set to true to enable tools that access user PII data (default: false) |
ITERABLE_ENABLE_WRITES |
No | Set to true to enable tools that perform write operations (default: false) |
ITERABLE_ENABLE_SENDS |
No | Set to true to enable tools that can send messages (default: false). Requires writes to be enabled |
- The CLI supports selecting the US or EU region, or entering a custom endpoint.
- For security, HTTPS is required for custom endpoints. The only exception is localhost development:
- Allowed:
http://localhost:3000,http://127.0.0.1:8080,http://[::1]:4000 - Not allowed: plain
http://on non-local hosts (usehttps://instead)
- Allowed:
- When a non-
*.iterable.comdomain is provided, the CLI will ask for confirmation.
- Claude CLI missing: install
claudeCLI, then re-runnpx @iterable/mcp setup --claude-code. - macOS Keychain issues: Ensure Keychain is accessible and re-run setup if needed.
Tool limit: Windsurf has a maximum limit of 100 tools that Cascade can access at any given time. When all permissions are enabled (ITERABLE_USER_PII=true, ITERABLE_ENABLE_WRITES=true, ITERABLE_ENABLE_SENDS=true), the Iterable MCP server exposes 104 tools, which exceeds this limit.
Workaround: Use restricted permissions to stay under the 100-tool limit:
- With default permissions (all disabled): 26 tools ✅
- With PII only: 37 tools ✅
- With PII + writes: 86 tools ✅
- With all permissions: 104 tools ❌ (exceeds Windsurf limit)
You can configure permissions when adding a key:
npx @iterable/mcp keys add --advancedOr update an existing key's permissions:
npx @iterable/mcp keys update <key-name> --advancedProcess persistence: After switching API keys with keys activate, you must fully restart Windsurf (quit and reopen the application). Windsurf keeps MCP server processes running in the background, and they don't automatically reload when you switch keys.
Tool limit: Antigravity has a maximum limit of 100 tools per MCP server. When all permissions are enabled (ITERABLE_USER_PII=true, ITERABLE_ENABLE_WRITES=true, ITERABLE_ENABLE_SENDS=true), the Iterable MCP server exposes 104 tools, which exceeds this limit.
Workaround: Use restricted permissions to stay under the 100-tool limit:
- With default permissions (all disabled): 26 tools ✅
- With PII only: 37 tools ✅
- With PII + writes: 86 tools ✅
- With all permissions: 104 tools ❌ (exceeds Antigravity limit)
You can configure permissions when adding a key:
npx @iterable/mcp keys add --advancedOr update an existing key's permissions:
npx @iterable/mcp keys update <key-name> --advancedSee CONTRIBUTING.md for development setup, building from source, and running tests.
Iterable's MCP server is currently in beta. MCP functionality may change, be suspended, or be discontinued at any time without notice. This software is provided "as is" and is open source and ready for you to experiment with. For more information, refer to Iterable Beta Terms.
This project is licensed under the MIT License - see the LICENSE file for details.

