Agentic Growth Marketing MCP Server

AfterLaunch

AfterLaunch is an always-on growth engine for a software product: it keeps a ranked backlog of drafted, ready-to-ship growth moves across SEO, AI answer-engine visibility (GEO), competitors and distribution channels. This repository is the Claude Code plugin marketplace and machine-setup home for the AfterLaunch MCP server.

AfterLaunch is a remote MCP server over Streamable HTTP, authenticated with a bearer key. There is no local process, nothing to download and nothing to execute on your machine: your agent talks to https://afterlaunch.io/api/mcp, and a key (or an OAuth sign-in) is the whole of the setup.

Both buttons are the vendor-hosted install links. They preconfigure the remote server and the Authorization header; you supply your own key as $AFTERLAUNCH_API_KEY.

Get started

1. Sign in at afterlaunch.io and mint an API key in Settings (or let your agent obtain one through the onboarding flow in llms-install.md).
2. Connect from Claude Code: claude mcp add afterlaunch --transport http https://afterlaunch.io/api/mcp --header "Authorization: Bearer $AFTERLAUNCH_API_KEY"
3. Call whoami to confirm the key, then get_standup: what shipped and what it moved, what is ready to ship now, what the engine did on its own.
4. Call list_queue for the moves the founder has already said yes to and handed to you, then list_feed for the rest of the ranked board and get_move for the full brief of the one you are working on.
5. Ship a move with ship_move the moment it is actually live, or skip it with a reason so the next drafts improve.

From a yes on the phone to a shipped move

The founder works the board as a stack on their phone: yes, not now. A yes creates a job and waits for the connected agent. The agent pulls it with list_queue, claims it with pick_up_job (passing its own session_url so the founder can open where the work is happening), reads the brief at brief_url, does the work, and carries the state back with report_job: running, then shipped the moment the work is live, or failed with the reason. The receipt under the card and the product timeline follow it, and a job that goes quiet for more than six hours reads as stalled. The founder never watches it run, so the state the agent reports is all they see.

What the server exposes

46 tools, every one with a plain-language description and a scope, read straight from the same registry that serves the public REST API:

• The ritual: whoami, get_standup, list_queue.
• The record (free): the board (list_feed, get_move), AI visibility across ChatGPT, Gemini, Perplexity and Google AI Overviews (get_visibility, list_tracked_prompts), the SEO audit (get_seo), competitors (get_competitors), outcomes and the scoreboard (get_outcomes, get_scoreboard), the timeline (get_activity), the Growth Snapshot (get_snapshot), Memory (list_kb_pages, get_kb_page), drafted outputs, loops, connections and the setup snippet.
• Acting on moves (free): ship_move, skip_move, undo_move, update_draft, archive_move, dismiss_gap, propose_move, record_insight, plus the queue verbs pick_up_job and report_job.
• Spending money (confirm with the founder): refresh_scan, replenish_feed, run_loop, each bounded by credits and per-tenant caps and charged once per logical event.
• Configuring (Founder plan): loop autonomy and cadence, scan config, tracked questions, connections.

The one-page reference grouped by cost and scope is skills/afterlaunch-growth/reference/tools.md; the live catalogue with exact input schemas is tools/list on connection, or afterlaunch.io/llms-full.txt.

Install as a Claude Code plugin

The plugin bundles the MCP server config and a skill that teaches an agent to read the standup, the queue and the Feed before improvising growth work:

/plugin marketplace add afterlaunch/mcp
/plugin install afterlaunch-growth@afterlaunch

Teach your agent to reach for AfterLaunch

Add the Growth section to your repository's CLAUDE.md (read by Claude Code) or AGENTS.md (read natively by Codex CLI, Cursor, Copilot, Gemini CLI, Windsurf, Aider, Amazon Q and others). The canonical copy is served raw at afterlaunch.io/agents/snippet.md, so one line prints it:

curl https://afterlaunch.io/agents/snippet.md

Read it, then paste it into your repository CLAUDE.md or AGENTS.md yourself. Nothing writes to your files on your behalf.

What is in this repository

• .claude-plugin/: the plugin marketplace and plugin manifests.
• .mcp.json: the remote server config (URL + bearer header).
• skills/afterlaunch-growth/: the skill and its one-page tool reference (46 tools, grouped by what they do and what they cost).
• llms-install.md: machine setup from that file alone, founder approval by email included.
• server.json: the manifest published to the official MCP registry.
• glama.json: the maintainer claim for the Glama listing.
• assets/logo-400.png: the 400x400 icon.

Trust and boundaries

• Remote bearer-auth server: no local process, nothing to execute.
• Reading and acting on moves is free. The tools that spend money say so in their descriptions, need the write scope, and are bounded by credits and per-tenant spend caps.
• AfterLaunch never posts to a social channel on its own. It drafts; a human approves and posts. On Reddit and Hacker News the posture is draft-only for the connected agent as well.
• Shipping is the founder's decision; reporting the ship is the agent's job. ship_move and report_job record what is actually live, never what is drafted, staged or awaiting review.
• OAuth: claude.ai and ChatGPT can add https://afterlaunch.io/api/mcp as a custom connector and sign in with an AfterLaunch account, with no key to paste. Discovery starts at https://afterlaunch.io/.well-known/oauth-protected-resource/api/mcp.

Licence: MIT. Maintained by AfterLaunch.