AfterLaunch: the agentic growth marketing engine

AfterLaunch MCP server

AfterLaunch watches how AI search and Google find your product, and hands you the next growth move with a draft ready. This repository is the home of its remote MCP server: connect Claude, ChatGPT, Claude Code, Codex or Cursor and your own AI can read the ranked moves, check AI visibility, take a move through to shipped and report it live.

• Server URL: https://afterlaunch.io/api/mcp (Streamable HTTP, nothing to install or run locally)
• Tools: 58 (see the list below)
• Sign-in: OAuth 2.0 for Claude and ChatGPT, or an API key for coding agents
• Cost: the free scan and the first 14 days cost nothing and need no card; then one plan, $149 a month
• Licence: MIT

The Cursor button opens Cursor's own confirmation screen with the server filled in. It reads your key from the AFTERLAUNCH_API_KEY environment variable, so no key is ever written into a link or a file.

What an agent can do with it

• Read the ranked growth moves. list_feed returns the backlog across SEO, AI visibility, competitors and distribution. get_move opens one with its evidence, source material, instructions and a prepared draft.
• Read AI visibility. get_visibility shows how ChatGPT, Gemini, Perplexity and Google AI Overviews answer the customer questions you track, with share of voice, the sources they cite and the gaps where rivals are named and you are not. get_seo, get_competitors and get_scoreboard cover the rest of the record.
• Take a move and ship it. accept_move, update_draft, then ship_move once the work is live. Moves handed to the agent arrive through list_queue, pick_up_job and report_job.
• Report it live. ship_move records what is live and a receipt follows on the timeline. Later, get_outcomes and get_scoreboard show what it earned.

A good first session is two calls: whoami to confirm which product the key drives, then get_standup for what shipped, what is ready now and what the engine did on its own.

Get started

Run this in your own terminal:

npx afterlaunch connect

It finds your client, takes your key or creates your account, writes the config and tells you to restart. It covers Claude Code, Codex, Cursor, Windsurf, Gemini CLI and VS Code, and the key never passes through a chat. The steps below are the manual path.

Install the plugin

The plugin bundles the server config and three skills: afterlaunch-growth, which teaches the agent to read the standup, the queue and the Feed before improvising growth work, plus roughen and findable from afterlaunch/skills. Create an API key first at afterlaunch.io/app/connect#keys, or run npx afterlaunch connect.

Install in Claude Code

/plugin marketplace add afterlaunch/mcp

/plugin install afterlaunch-growth@afterlaunch

When you enable the plugin, Claude Code asks for your AfterLaunch API key. It masks the field and keeps the key in your system's secure credential store, not in a settings file, and sends it only as the bearer header to https://afterlaunch.io/api/mcp.

Install in Cursor

Install AfterLaunch Growth from the Cursor Marketplace, then set AFTERLAUNCH_API_KEY under Plugins, Configure on the plugin. The plugin's mcp.json holds only the ${AFTERLAUNCH_API_KEY} placeholder, never the key. To use the server without the plugin, use the install button above or the manual config under Cursor below.

For the manual configs below (Cursor and Codex), export the key in your shell before you open the client:

export AFTERLAUNCH_API_KEY=al_your_key_here

Connect in one minute

You need a scanned product first. Run the free scan at afterlaunch.io, because connecting needs a scanned product. The server URL is the same everywhere:

https://afterlaunch.io/api/mcp

Never put a key in that URL, and never paste a key into a chat.

Claude (web and mobile)

1. Open Customize, then Connectors, and add a custom connector.
2. Paste https://afterlaunch.io/api/mcp as the server URL and press Add.
3. Open AfterLaunch under Connectors, press Connect, and approve.

You sign in with your AfterLaunch account, so there is no key to create. Add it once on claude.ai and it is available in the mobile app. Free Claude plans allow one custom connector. On a Team or Enterprise plan an owner adds it once in organisation settings.

ChatGPT

1. Open Settings, then Security and login, and turn on Developer mode.
2. In Plugins, press the plus button to create a developer-mode app.
3. Paste https://afterlaunch.io/api/mcp as the app's endpoint.
4. Choose OAuth for authentication and create the app.
5. Sign in to AfterLaunch and approve.

Claude Code

Create an API key at afterlaunch.io/app/connect#keys, then run:

claude mcp add afterlaunch --scope user --transport http https://afterlaunch.io/api/mcp --header "Authorization: Bearer YOUR_AFTERLAUNCH_API_KEY"

Replace YOUR_AFTERLAUNCH_API_KEY with the key. It goes into your private user config (~/.claude.json), outside every repository, so it is never committed. Or install the plugin, which asks for the key itself.

Codex

codex mcp add afterlaunch --url https://afterlaunch.io/api/mcp --bearer-token-env-var AFTERLAUNCH_API_KEY

Codex reads the key from AFTERLAUNCH_API_KEY in your environment and sends it as a bearer header. The token never lands on disk.

Cursor

Use the install button above, or add this to .cursor/mcp.json (per project) or ~/.cursor/mcp.json (every project):

{

"mcpServers": { "afterlaunch": { "url": "https://afterlaunch.io/api/mcp", "headers": { "Authorization": "Bearer ${env:AFTERLAUNCH_API_KEY}" } } } }

Cursor expands the placeholder from your shell, so set AFTERLAUNCH_API_KEY before you open the project.

Other clients

The onboarding guide at afterlaunch.io/agents/onboard.md carries the exact command or config for VS Code, Windsurf, Gemini CLI and Cline too, and says which have been checked. It is written for an AI to read, so you can also just tell your agent to follow it. No account yet? An agent that can run commands can start one for you: it sends your site address and email, you approve one email, and the key is handed over once. The steps are in that same guide.

Always-on agents

An agent that runs on a schedule (OpenClaw, Hermes Agent, a Claude scheduled task, the Claude Agent SDK) uses the same server and the same key. Each morning it calls get_standup and relays the screen to you in its own chat channel, Telegram, Slack or wherever you talk to it. When you name a move, it accepts it, wins the job with pick_up_job, asks at most one question through report_job, and reports the move shipped with its live address only after you have said yes to publishing it.

The standing orders for that agent are served raw, so one line prints them:

curl https://afterlaunch.io/agents/always-on.md

Paste them where your runtime keeps its standing instructions. The per-runtime connect lines, scheduler lines and where each one keeps its orders are at afterlaunch.io/agents/always-on.

Gemini CLI

gemini extensions install https://github.com/afterlaunch/mcp

Gemini CLI asks for your AfterLaunch API key and keeps it in the system keychain. It sends it only as the bearer header to https://afterlaunch.io/api/mcp.

Tools

58 tools. The live catalogue with exact input schemas is tools/list on connection, and the one-page reference in this repository is skills/afterlaunch-growth/reference/tools.md.

Scopes: 26 tools need read, 26 need act, 3 need write and 3 need config.

Start a session

Read the ranked growth moves

Read AI visibility, SEO and the record

Take a move and ship it

| list_queue | act | Moves the owner has said yes to and handed to the agent, o