Send emails from natural language prompts.
https://next.hal9.ai/mcp/mcpbuild/send-email
First time connecting an MCP? See Installing MCPs for setup steps in Claude, Claude Code, and other clients.
send-email turns a plain-language request into a real email. Instead of filling out a rigid "to / subject / body" form, an agent (or a human) just describes what they want sent, and the MCP figures out the rest:
send email to javier@hal9.ai with text hello!
Under the hood, the prompt is passed to a language model with a single
send_email tool definition. The model extracts the recipient,
infers a short subject when one isn't given, and writes the body — then the
MCP calls the email provider (Resend) to
actually send it.
Once connected, an agent sees a single tool — send_email — with
three parameters:
| Parameter | Description |
|---|---|
| to | Recipient email address |
| subject | Subject line — inferred from the request if not stated |
| text | Plain-text body of the email |
The agent doesn't need to call the tool directly with structured arguments — it can just be asked in plain English, and it will:
Typical uses: notifying someone from an automated workflow, letting an agent follow up on a task by email, or wiring "email X" into a larger multi-step agent without writing any email-sending code.
send-email is open source and built on the Hal9 agent structure. You can run it yourself instead of using the hosted version:
cd send-email pip install -r requirements.txt export GROQ_API_KEY=… export RESEND_API_KEY=… echo "send email to you@example.com with text hello!" | python app.py
| Variable | Required |
|---|---|
| GROQ_API_KEY | Yes |
| RESEND_API_KEY | Yes |
| RESEND_FROM | No — default sender address |
| GROQ_MODEL | No — defaults to qwen/qwen3.6-27b |
Full source, deploy workflow, and contributing guide: github.com/hal9ai/mcp.build/tree/main/send-email
Connecting from Claude, Claude Code, Cursor, or another MCP client? Point it at the endpoint above — full step-by-step instructions live on the Installing MCPs page.