For AI agents · MCP
Give your agent a mailbox.
MailStrap ships a native MCP server. Connect Claude, Cursor, or any Model Context Protocol client with your API key, and let your agent read threads, send mail, manage DNS, grow audiences and check campaign stats - scoped to your organization, metered and audit-logged like every other send.
12 org-scoped tools · streamable HTTP · works with any MCP client
# Connect Claude Code to your workspace
claude mcp add --transport http mailstrap \ https://api.mailstrap.com/v1/mcp \ --header "Authorization: Bearer mf_live_..."
# Then just ask:
> “Check hello@acme.com for unread invoices and
draft replies to each.”
Connected in two minutes
Create an API key
One key per organization from Settings. It scopes every tool call to your workspace, nothing else.
Add the MCP server
Point any MCP client at api.mailstrap.com/v1/mcp with your key as a bearer header. Claude Code, Claude.ai, Cursor and n8n all speak it.
Put your agent to work
Triage inboxes, draft and send replies, verify domain DNS, add subscribers, pull campaign stats - the same actions your dashboard can take.
12 tools, the same guardrails as the dashboard
- read_thread and list_messages: full inbox access per mailbox
- send_email: quota-checked, metered, suppression-aware
- get_domain_dns and verify_domain: SPF, DKIM, DMARC status
- add_subscriber: fires your existing automations
- campaign_stats: delivery, open and click rates
- usage_overview: plan and consumption at a glance
- Every write is audit-logged with the acting key
- Org-scoped by design: a key only sees its own workspace