
One command in the terminal and Claude Code holds the whole mail room — it sends, reads what comes back, replies in-thread, and routes anything big to you for approval.
Free sandbox to integrate · Production from $3/month
Claude Code registers MCP servers from the command line. Scope a key in the dashboard, run the command, restart the session — the tools are just there.
Full setup guide →$ claude mcp add drin \
-e DRIN_API_KEY=drin_k_… \
-- npx -y @drin00/mcp
# restart the session — 54 tools appear.
# also works in the Claude desktop app via mcpServers.Three real tasks, end to end — copy a prompt and watch it run the tools.
Create an inbox support@mydomain.com and a webhook to https://myapp.com/inbound, then simulate an inbound email and show me the parsed payload.
Send a test email from hello@mydomain.com to my address with subject "Drin wired up", then check whether it delivered and show me the event trail.
Draft a campaign to the NEWS topic announcing v2.0 — pull the highlights from the changelog — and route it to me for approval. Don't send anything yourself.
Both. Claude Code adds it with the one-liner; the desktop app takes the same server in its mcpServers config. Any MCP client gets the identical 54 tools.
Scope the key in the dashboard and start in the sandbox, where sends only reach verified recipients. Anything wide — a campaign, an automation going live — waits for a human reply regardless of what the agent wants. How approve-by-reply works →
Everything the dashboard can: send and batch mail, create inboxes, read threads and reply in them, verify domains, manage templates, webhooks and suppressions, pull metrics, and draft campaigns.
No — the MCP server alone is enough. The skills add worked procedures (compose, triage an inbox, verify a domain, report metrics) so Claude follows the good path instead of improvising one.
No. MCP, the CLI and skills ship on every plan, including the free sandbox. Pricing →