Summit Server

npm

Bring Summit's conversion-audit insights into your coding agent. Hand Claude Code / Codex / Gemini CLI a Summit audit and have it implement the fixes — grounded in real CRO analysis with exact selectors and before→after copy.

Install

Nothing to install — any MCP client can launch it straight from npm:

npx -y @summit-ai/summit-mcp

Or install the summit-mcp command globally:

npm install -g @summit-ai/summit-mcp

Requires Node.js ≥ 18.17.

Tools

The first four work off a public share token — no auth required. The workspace tools need SUMMIT_API_TOKEN + SUMMIT_WORKSPACE_ID and cover the full Study → Approve → Ship loop, so an agent can go from audit to launched experiment to measured result without leaving the editor. Tools marked mutates change workspace state (they never touch your live site directly — variants only serve after an experiment is explicitly launched).

Getting the two workspace values (a Pro feature — minting a token requires a paid plan): sign in and open Settings → Summit MCP tokens in the app to generate SUMMIT_API_TOKEN (shown once) and copy your SUMMIT_WORKSPACE_ID. The web docs at trysummit.ai/docs walk through it and pre-fill the config with your workspace ID. A token is scoped to a single workspace and can be revoked anytime.

Configuration (env)

Register with your agent

The audit tools work with just SUMMIT_API_BASE_URL. To unlock the workspace loop, add SUMMIT_API_TOKEN + SUMMIT_WORKSPACE_ID (from Settings → Summit MCP tokens; see above).

Claude Code

claude mcp add summit \
  --env SUMMIT_API_BASE_URL=https://api.trysummit.ai \
  --env SUMMIT_API_TOKEN=smt_your_token_here \
  --env SUMMIT_WORKSPACE_ID=your_workspace_id \
  -- npx -y @summit-ai/summit-mcp

Codex CLI — ~/.codex/config.toml

[mcp_servers.summit]
command = "npx"
args = ["-y", "@summit-ai/summit-mcp"]
env = { SUMMIT_API_BASE_URL = "https://api.trysummit.ai", SUMMIT_API_TOKEN = "smt_your_token_here", SUMMIT_WORKSPACE_ID = "your_workspace_id" }

Gemini CLI — ~/.gemini/settings.json

{
  "mcpServers": {
    "summit": {
      "command": "npx",
      "args": ["-y", "@summit-ai/summit-mcp"],
      "env": {
        "SUMMIT_API_BASE_URL": "https://api.trysummit.ai",
        "SUMMIT_API_TOKEN": "smt_your_token_here",
        "SUMMIT_WORKSPACE_ID": "your_workspace_id"
      }
    }
  }
}

(Any MCP-aware client works — point it at npx -y @summit-ai/summit-mcp, or at the summit-mcp command if installed globally, over stdio. Omit the token + workspace-id envs to use just the free audit tools.)

Example agent flow

You: Audit https://moonsign.co.in and fix the top 3 conversion issues.

Agent → summit_run_audit(url="https://moonsign.co.in", email="[email protected]")
      ← { share_url: ".../audit?r=TOK", status: "queued" }
Agent → summit_implementation_plan(report="TOK")     # poll until completed
      ← { steps: [ { selector: "a.cta", before: "Learn more",
                     after: "Get my free reading", expected_lift_pct: 12 }, … ] }
Agent then edits the codebase per each step.

Or drive the whole loop against your workspace (auth env vars set):

You: Anything waiting on me? Approve the highest-lift fix and launch it.

Agent → summit_review_queue()
      ← { findings: [ { id: "…", title: "Weak hero CTA", estimated_lift: 14, … } ], experiments: [] }
Agent → summit_approve_finding(finding_id="…")        # builds experiment + variants
Agent → summit_approve_experiment(experiment_id="…")  # after QA
Agent → summit_launch_experiment(experiment_id="…")
      … later …
Agent → summit_experiment_results(experiment_id="…")
      ← { winner_key: "b", leader_prob: 0.97, is_significant: true }

Develop

npm install
npm test          # node:test — pure transforms + mocked-fetch tool tests, no network

The package is plain ESM JavaScript with no build step. Runtime dependencies are just @modelcontextprotocol/sdk and zod.

License

MIT