Email infrastructure for AI agents.

SendCore connects AI agents like Claude, Cursor, and Copilot directly to your email infrastructure. Send emails, check delivery logs, and give agents their own inboxes — all through natural language, the MCP server, or the REST API.

$npx sendcore-mcp

How it works

Connect your AI agent to SendCore via MCP or the REST API — no manual API calls needed.

01

Generate Key

Create an API key from your SendCore dashboard.

02

Choose Your Connection

Use MCP (Claude, Cursor), the REST API with function calling (OpenAI, LangChain), or the Node.js and Python SDKs.

03

Ask in Natural Language

Tell your AI agent what to do — "Send a welcome email to new subscribers"

04

Agent Executes

The AI agent calls SendCore APIs directly. No code to write.

8 tools at your agent's disposal

Core SendCore features are exposed via MCP — send emails, check logs, and manage agent inboxes.

Email

  • send_email

    Send transactional emails with HTML or plain text

  • email_logs

    Check recent delivery status and timestamps

Agent Inboxes

  • create_inbox

    Create an email inbox for your AI agent

  • list_inboxes

    View all agent inboxes and their addresses

  • read_emails

    Read inbox emails with semantic search

Inbox Actions

  • get_email

    Fetch a single inbound email in full

  • extract_otp

    Pull verification codes out of inbound emails

  • send_as_agent

    Send or reply from an agent inbox address

Claude Setup

Connect via the hosted gateway with OAuth (no API key needed), or run locally with stdio.

Full documentation

Option A — Hosted (OAuth)

{
  "mcpServers": {
    "sendcore": {
      "type": "http",
      "url": "https://mcp.usesendcore.com/mcp/v1/protocol"
    }
  }
}

Option B — Local (API key)

{
  "mcpServers": {
    "sendcore": {
      "command": "npx",
      "args": ["-y", "sendcore-mcp"],
      "env": {
        "SENDCORE_API_KEY": "sc_mcp_your_key_here"
      }
    }
  }
}

Online MCP Inspector

Test the MCP server from your browser using the MCP Inspector or any MCP-compatible HTTP client.

Full documentation

Connect to our hosted gateway at https://mcp.usesendcore.com using the Streamable HTTP transport:

npx @modelcontextprotocol/inspector \
  --transport http \
  --url https://mcp.usesendcore.com/mcp/v1/protocol \
  --header "Authorization: Bearer sc_mcp_your_key_here"

REST API

Use the SendCore REST API directly with any AI framework that supports function calling.

OpenAI Function Calling

{
  "tools": [{
    "type": "function",
    "function": {
      "name": "send_email",
      "description": "Send a transactional email",
      "parameters": {
        "type": "object",
        "properties": {
          "from": { "type": "string" },
          "to": { "type": "array", "items": { "type": "string" } },
          "subject": { "type": "string" },
          "html": { "type": "string" }
        },
        "required": ["from", "to", "subject", "html"]
      }
    }
  }]
}

REST API Quick Example

# Any AI agent can call SendCore over HTTP
curl -X POST https://api.usesendcore.com/api/v1/emails/send \
  -H "x-api-key: sc_live_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "from": "hello@yourdomain.com",
    "to": ["user@example.com"],
    "subject": "Hello from AI",
    "html": "<h1>Hello World</h1>"
  }'

Example prompts to try

"Send a welcome email to user@example.com with subject Welcome! and HTML body '<h1>Hello!</h1>'"

"What are my recent email logs? Show me the last 10 deliveries."

"Show me any failed deliveries in my recent email logs."

"Create an inbox for my signup-testing agent."

"List my agent inboxes and their addresses."

"Find the email about a refund in my support inbox."

"Extract the OTP code from the newest email in my agent inbox."

"Reply to the latest email in my support inbox and say we're on it."

Ready to let AI handle your email infrastructure?

Get started free — no credit card required. Generate your API key and connect your AI agent in minutes.