How many tools does it actually expose?

Ten today. The package registers more, but the surface is revealed in waves and the committed default is the core wave plus the six always-on auth tools. You can opt into more with INTEMPT_MCP_TOOLS, and the enabled set is read-only until the write wave ships.

Do I need an API key?

No. Run the login tool in the chat and a browser opens on your normal Intempt sign-in, using authorization code with PKCE over a loopback redirect. The token refreshes itself.

What is the difference between the MCP server and the CLI?

The MCP server reads your live platform data from inside an agent conversation. The CLI is a build-time tool that scans your codebase and generates typed SDK wrappers. Different jobs.

MCP server · v1.2.0

Read your customer data from inside Claude Code.

Your segments, users and attributes, readable from inside the agent you already use. One command to add, browser login, no API key.

Start for free

Run this in your terminal

# Adds the Intempt MCP server to Claude Code

$ claude mcp add intempt -- npx -y @intempt-technologies/mcp

# Confirm it worked

$ claude mcp list

→ intempt appears in the list

Then run the login tool in the chat. A browser opens on your normal Intempt sign-in.

Not listed? The command must be run in the same shell profile Claude Code launches from. Node 18+ is required.

0.0on G2

1B+events

7+years profitable

50+companies

How it works

Describe the outcome. It calls the right tool.

Add the server

One command. Claude Code, Claude Desktop, Cursor, Windsurf, VS Code, Zed and Cline all speak MCP, so the same server works in whichever one you already have open.

Log in from the chat

Run the login tool. A browser opens on your normal Intempt sign-in, hands back a token over a loopback redirect, and every tool works from then on. No key to paste, nothing to store in a dotfile.

Ask in plain language

You describe the outcome and the nine skills route it to the right tool, rather than you picking an endpoint.

It reads the real platform

Tools hit the same REST APIs the console uses, so what the agent reads is your live data, not an export. Today the enabled surface is read-only.

Skills

Nine skills, installed with the server.

A skill is a procedure the agent follows instead of guessing which tool to reach for. They land in ~/.claude/skills on install.

/instrument

Scan the codebase, propose a tracking plan, generate typed wrappers

/analyze

Funnels, retention and paths without opening the console

/market

Journeys, personalization and experiments from the prompt

/sell

Pipeline, meetings and follow-ups against the live profile

/design

Brand-kit-aware images, video, pages and messages

/crm

Look up and update people and accounts by name, not by id

/registry

Discover which tools are enabled and what each can do

/setup

Org, project and source wiring on first run

/intempt

The router - picks the skill when you have not named one

CLI

Instrument the codebase, not a spreadsheet.

The MCP server operates the platform. The CLI instruments your app. It reads your source, proposes the events worth tracking, writes them to intempt.yaml, and generates typed wrappers for 14 platforms so an analytics call that does not match the plan fails at compile time.

See the CLI

$ npm install -g @intempt-technologies/cli

$ intempt init
  → scans your source
  → proposes events + properties
  → writes intempt.yaml

$ intempt generate
  → typed wrappers for your platform

$ intempt status
  → instrumented vs missing

Frequently asked questions

Everything you need to know about the Intempt MCP server and CLI.

Claude Code, Claude Desktop, Cursor, Windsurf, VS Code with Copilot, Zed, Cline, and anything else that speaks MCP. It ships both a stdio and an HTTP transport, so a desktop client and a remote one both work.

Add the server. Ask it something.

Free to install, works with whichever MCP client you already have open, and logs in through the browser you are already signed into.

Start for free