SuperSend TX MCP server

Expose SuperSend TX to Cursor, Claude Code, or any MCP client with a curated tool surface.

Package: supersendtx-mcp on npm
Version: 0.6.0

Docs index for agents: llms.txt · Agent Skills · Agent skill notes

For Ranla branding, use ranla-mcp (npx ranla-mcp) with RANLA_API_KEY=rnl_…. Arc tools call the app host (app.ranla.ai); mail tools call api.ranla.ai.

Transports

Local HTTP listens on http://127.0.0.1:3000/mcp by default (--host / --port override). GET /health is unauthenticated (local and hosted).

Hosted OAuth (recommended)

Remote MCP clients (Cursor, Claude Code) can connect without pasting an API key:

1. Add the hosted URL only — https://mcp.supersendtx.com/mcp (no headers block).
2. The client discovers OAuth via WWW-Authenticate + /.well-known/oauth-protected-resource.
3. Sign in on app.supersendtx.com, approve access, and the client receives tokens.

Authorization server metadata: https://app.supersendtx.com/.well-known/oauth-authorization-server

Bearer stx_… still works for scripts, CI, and advanced setups.

Install in Cursor (stdio)

1. Create an API key in the dashboard (or use Get started → Copy setup prompt / Install in Cursor).
2. Add to .cursor/mcp.json (or global ~/.cursor/mcp.json):

{
  "mcpServers": {
    "supersendtx": {
      "command": "npx",
      "args": ["-y", "supersendtx-mcp"],
      "env": {
        "SUPERSENDTX_API_KEY": "stx_your_key_here"
      }
    }
  }
}

3. Restart Cursor. The tools appear in Agent chat.

Optional: set SUPERSENDTX_API_URL when pointing at a local app (http://localhost:3003/api).

The dashboard Install in Cursor button generates a deeplink with your key pre-filled.

Cursor (HTTP)

Terminal:

npx -y supersendtx-mcp --http --port 3000

.cursor/mcp.json (URL + Bearer; exact shape depends on Cursor version):

{
  "mcpServers": {
    "supersendtx": {
      "url": "http://127.0.0.1:3000/mcp",
      "headers": {
        "Authorization": "Bearer stx_your_key_here"
      }
    }
  }
}

Replit (hosted remote)

One-click install of the hosted server (OAuth — no API key in the link):

App-side sends in a Replit project still use Secrets + code — see Replit builder guide.

Cursor (hosted remote)

OAuth (recommended): no local process, no API key in config.

{
  "mcpServers": {
    "supersendtx": {
      "url": "https://mcp.supersendtx.com/mcp"
    }
  }
}

Cursor runs the browser OAuth flow on first connect.

Bearer (advanced): dashboard Copy hosted MCP JSON pre-fills your key.

{
  "mcpServers": {
    "supersendtx": {
      "url": "https://mcp.supersendtx.com/mcp",
      "headers": {
        "Authorization": "Bearer stx_your_key_here"
      }
    }
  }
}

Health check: GET https://mcp.supersendtx.com/health

Claude Code

# stdio
claude mcp add --transport stdio supersendtx -- npx -y supersendtx-mcp

# HTTP (local — start --http server first)
claude mcp add --transport http supersendtx http://127.0.0.1:3000/mcp

# HTTP (hosted)
claude mcp add --transport http supersendtx https://mcp.supersendtx.com/mcp

Export SUPERSENDTX_API_KEY=stx_… for stdio. For HTTP (local or hosted), configure Bearer headers in the client.

Tools

All tools call the supersendtx SDK (no duplicate HTTP).

Arc (growth agent)

Talk to Arc from Claude / Cursor without opening the dashboard. Mutations go through Arc; read tools are talk-mode only.

Example flow: arc_ensure_thread → arc_message (“draft a re-activation campaign”) → if gated, arc_list_approvals → arc_approve.

Set SUPERSENDTX_APP_URL (or RANLA_APP_URL) when the app host is not the default. Do not point Arc at the transactional API host.

send_email

Inputs: from, to, subject, optional html / text / reply_to.

Sandbox: if from is [email protected], recipients must match the account email.

send_test_webhook_event

Inputs:

• event (required) — e.g. email.bounced, email.delivered
• optional email_id (msg_…)
• optional deliver (default true) — enqueue to subscribed webhooks

Install

npx -y supersendtx-mcp
npx -y supersendtx-mcp --http --port 3000
npx -y ranla-mcp
# or
npm install -g supersendtx-mcp

For agents without MCP, use the dashboard Copy setup prompt flow, Agent Skills, or agent-skill.md with llms.txt.

The package is also listed on Smithery for agent discovery.