launchrank

CLI and MCP server for LaunchRank — run AI search-visibility audits from your terminal or from an AI agent. An audit scores how a product shows up across AI answer surfaces and returns visibility gaps, opportunities, and growth actions.

Install

npm i -g launchrank      # global install
# or run without installing:
npx -y launchrank audit https://example.com

Getting an API key

1. Sign in at getlaunchrank.com.
2. Go to Settings → API keys.
3. Create a key (looks like lr_...) and set it as LAUNCHRANK_API_KEY.

Without a key the CLI prints a local mock audit so you can see the output shape.

Audit usage

export LAUNCHRANK_API_KEY=lr_xxx

launchrank audit https://example.com                     # markdown to stdout
launchrank audit example.com --format json               # JSON to stdout
launchrank audit https://example.com --out audit.md      # write to a file
launchrank audit https://example.com --dry-run           # local mock, no API call

Options:

| Option | Description | | --- | --- | | --format markdown\|json | Output format (default markdown). | | --out <file> | Write output to a file instead of stdout. | | --api-url <url> | Override the app base URL (default https://getlaunchrank.com). | | --token <value> | API key / bearer token. | | --dry-run | Print a local mock audit without calling the API. |

Token resolution: --token → LAUNCHRANK_API_KEY → LAUNCHRANK_AGENT_TOKEN. Exit codes: 0 success, 1 usage error, 2 API error.

Find customers

launchrank scout https://example.com                                  # ranked prospect targets, markdown
launchrank scout example.com --audience "solo SaaS founders" --format json
launchrank scout example.com --category "revenue analytics" --competitors "PostHog,Mixpanel"

Returns ranked customer targets (fit reason, proof signal, outreach angle, ready-to-send opener) plus search queries across Product Hunt, TrustMRR, X, and Reddit. Requires an API key (no mock).

MCP server

launchrank mcp starts a stdio MCP server exposing three tools:

• run_audit — { url } → structured audit JSON (visibilityScore, summary, promptGaps, opportunities, growthActions).
• audit_markdown — { url } → the audit as a markdown report.
• find_customers — { url, category?, audience?, competitors? } → ranked prospect targets with openers and search queries.

Both read the key from LAUNCHRANK_API_KEY (or LAUNCHRANK_AGENT_TOKEN). There is no mock in MCP — without a key the tools return an error.

Claude Code

claude mcp add launchrank -e LAUNCHRANK_API_KEY=lr_xxx -- npx -y launchrank mcp

Cursor / Windsurf

Add to your MCP config JSON:

{
  "mcpServers": {
    "launchrank": {
      "command": "npx",
      "args": ["-y", "launchrank", "mcp"],
      "env": {
        "LAUNCHRANK_API_KEY": "lr_xxx"
      }
    }
  }
}

License

MIT