@molted/mcp

MCP server for the Molted Email platform. Exposes the Molted API as discoverable tools for any MCP-compatible agent runtime (Claude Desktop, Cursor, VS Code Copilot, Windsurf, and others).

Quick start

npx -y @molted/mcp

Set MOLTED_API_KEY to your API key (mm_live_* or mm_test_*). The server starts in stdio mode by default.

Configuration

Environment variables

| Variable | Required | Default | Description | |----------|----------|---------|-------------| | MOLTED_API_KEY | Yes | - | Your Molted API key | | MOLTED_API_URL | No | https://api.molted.email | API base URL |

CLI options

--transport <stdio|http>  Transport mode (default: stdio)
--port <number>           HTTP port (default: 3100, http mode only)
--help                    Show help

Runtime configuration

Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "molted": {
      "command": "npx",
      "args": ["-y", "@molted/mcp"],
      "env": {
        "MOLTED_API_KEY": "mm_live_..."
      }
    }
  }
}

Cursor

Add to .cursor/mcp.json:

{
  "mcpServers": {
    "molted": {
      "command": "npx",
      "args": ["-y", "@molted/mcp"],
      "env": {
        "MOLTED_API_KEY": "mm_live_..."
      }
    }
  }
}

VS Code Copilot

Add to .vscode/mcp.json:

{
  "servers": {
    "molted": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@molted/mcp"],
      "env": {
        "MOLTED_API_KEY": "mm_live_..."
      }
    }
  }
}

Windsurf

Add to ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "molted": {
      "command": "npx",
      "args": ["-y", "@molted/mcp"],
      "env": {
        "MOLTED_API_KEY": "mm_live_..."
      }
    }
  }
}

HTTP transport (cloud agents)

MOLTED_API_KEY=mm_live_... npx -y @molted/mcp --transport http --port 3100

Available tools

Budget

| Tool | Description | |------|-------------| | check_budget | Check remaining send capacity (daily/hourly limits, bounce/complaint budget) |

Inbox & Threads

| Tool | Description | |------|-------------| | read_inbox | List threads with optional filtering by mailbox, status, and limit | | get_thread | Get a full thread with all messages and metadata | | reply | Reply to a thread (shorthand text/html or custom template) | | archive | Archive a resolved thread | | update_thread | Update thread status, metadata, or assignment | | approve | Approve a pending send in the approval queue |

Sending

| Tool | Description | |------|-------------| | send_email | Send a policy-checked email (shorthand or template mode) | | draft_email | Draft an email with AI composition and policy pre-check | | dry_run | Simulate a send to preview the policy decision | | batch_send | Send to multiple recipients with per-recipient policy (max 500) | | schedule_followup | Schedule a delayed follow-up that auto-cancels on reply |

Intelligence

| Tool | Description | |------|-------------| | classify_inbound | Classify an inbound email's intent and safety | | next_best_action | Get a recommendation for what to do next with a contact | | get_context | Get full contact timeline and relationship history | | batch_classify | Classify multiple inbound emails in one call | | batch_next_action | Get next-best-action for multiple contacts at once |

Scoped keys

API keys can be scoped to specific mailboxes. When using a scoped key, tools automatically operate within that mailbox's context. See the authentication docs for details.

Error handling

The MCP server maps API responses to structured MCP results:

• 2xx - Success with parsed data
• 429 - Rate limited (includes retry-after)
• 403 - Insufficient scope (key doesn't have required permissions)
• 401 - Authentication error
• 404 - Resource not found
• 5xx - Server error

Policy blocks (e.g., suppressed contact, rate limit exceeded) are returned as successful results with the block reason so your agent can reason about alternatives.

Docs

Full documentation: docs.molted.email/mcp-server

License

MIT