AI-Native Email Intelligence for Agents & IDEs
Connect your AI workflows to cmercury's high-performance transactional email engine and email verification using the Unified HTTP Streaming standard.
The cmercury MCP Server acts as a secure, standardized bridge between your AI agents and our enterprise email delivery engine. Whether you are using Cursor to debug transactional workflows or Claude to automate customer outreach, this server allows your AI to understand your account's health and trigger high-deliverability emails using natural language commands.
Pre-Flight Checklist: Get cmercury Ready
Signup & Verify Domain
Ensure the domain you intend to send from is already verified in your cmercury Dashboard.
Setup Automation Campaigns
Create your Automation Campaigns one time, based on the type of transactional mails you wish to use. These act as the logic for your transactional activities.
Get API Token
Navigate to Account Settings in cmercury to get/create your secure API token for MCP authentication.
Get TriggerIDs
Locate the TriggerID for your automation campaigns. You will provide these to the AI when using the sending tools.
Capabilities & Tools
Account & Status
-
check_connection Validate API token connectivity and bridge status.
-
get_email_credits Real-time check of prepaid email sending balance.
Verification
-
verify_single_email Determine if an email is Usable, Trash, or Neutral.
-
get_verify_status Account health and credits for cleaning services.
FireTrigger Delivery
-
send_transactional_email Standard API-driven transactional dispatch.
-
send_dynamic_template_email Merge personalized fields into dynamic templates.
-
send_with_attachments Robust support for CC, BCC, and file attachments.
Prompt Library
Copy these "Agent Instructions" to execute complex cmercury workflows. These are optimized for Cursor, Claude, and GitHub Copilot.
Full Credit & Connection Audit
Run check_connection, get_email_credits, and get_verify_status. Provide a complete summary of my remaining resources.
Real-time Lead Verification
"Verify the email [EMAIL] via cmercury. Only proceed if the status is 'Usable'. If it is 'Trash', explain why."
Personalized Template Dispatch
Trigger a dynamic template send to [EMAIL] using template [ID] with personalized merge fields.
Send with CC/BCC Support
Send an email to [TO_EMAIL] while keeping [CC_EMAIL] in the loop via send_with_attachments.
Verify Account Status Check
"Check my verification credits. If I have fewer than 1000, alert me to recharge."
Clean a List of Leads
"I have a list of emails [LIST]. Use verify_single_email for each and give me a table of Usable vs Trash."
Trigger FireTrigger Send
"Send a standard transactional welcome email to the user who just signed up at [EMAIL]."
Send PDF with CC
"Send the generated PDF to [CLIENT] and BCC my archive email using cmercury."
Connection Debugger
"Validate my token connection. If it fails, check if I'm hitting rate limits."
Threshold Credit Alert
"Check my credits. Compare them to last week's balance of [NUMBER] and tell me my burn rate."
Security & Token Management
Your cmercury API Token is a sensitive credential. Modern MCP clients allow you to inject this securely without exposing it in your codebase or chat history.
1. System Env Variable
The "Gold Standard" for security. Store your token as CMERCURY_API_TOKEN in your OS environment for seamless auto-detection.
2. Claude Vault
Use the "Authorization" or "Headers" field in the Claude Desktop UI. Map X-Token directly to your key; Claude handles the encryption.
3. Config References
Use environment references like ${env:VAR_NAME} in your settings.json to prevent leaking keys when sharing your config files.
Agent & Framework Guides
Cursor & Cline
Cursor and Cline (formerly Claude Dev) offer native MCP support. Simply add the endpoint to your MCP settings. The agent will automatically discover all 7 cmercury tools.
MCP Config (mcp.json)
{
"mcpServers": {
"cmercury": {
"url": "https://mcp.cmercury.com/mcp",
"env": { "CMERCURY_API_TOKEN": "${env:CMERCURY_API_TOKEN}" }
}
}
}LangChain & CrewAI
Use the official 2026 langchain-mcp-adapters to turn cmercury tools into native objects for your multi-agent "Crews".
from langchain_mcp_adapters.client import MultiServerMCPClient
from crewai import Agent, Task, Crew
client = MultiServerMCPClient({
"cmercury": {"url": "https://mcp.cmercury.com/mcp", "transport": "http"}
})
# Automatically discover and load all 7 tools
cmercury_tools = await client.get_tools()
email_agent = Agent(
role='Deliverability Specialist',
goal='Verify leads and send templates',
tools=cmercury_tools
)PydanticAI
For high-performance Python agents. Connect using the standard mcp Python SDK.
from mcp import ClientSession
from pydantic_ai import Agent
async with streamable_http_client("https://mcp.cmercury.com/mcp") as (read, write):
async with ClientSession(read, write) as session:
await session.initialize()
# Bind cmercury tools to a PydanticAI Agent
agent = Agent('openai:gpt-4o', deps_type=MyDeps)
# ... agent logic ...Jupyter & Google Colab
Perfect for bulk email verification or data cleaning. Use the Python MCP SDK to create a persistent connection within your notebook session.
# !pip install mcp
from mcp import ClientSession, streamable_http_client
async with streamable_http_client("https://mcp.cmercury.com/mcp") as (read, write):
async with ClientSession(read, write) as session:
await session.initialize()
# Example: Verify a single email from a dataframe
result = await session.call_tool("verify_single_email", {
"email": "lead@company.com",
"apiToken": "YOUR_TOKEN"
})
print(f"Status: {result.content[0].text}")Codex & Open Interpreter
Grant your local interpreter "Email Skills." When Codex runs locally, it can query cmercury to handle communication tasks or verify system alerts.
Interpreter Command
Once connected, simply ask: "Interpreter, use cmercury to check my credits."
Unlock Power User Tools
Beyond simple sending, our MCP server supports send_with_attachments for complex reporting and get_verify_status for high-volume list cleaning monitoring.