What we ship that your agent can’t build.
Your agent handles the reply. We handle the security and identity layer underneath — per-agent addresses, injection scanning, reputation tracking, MCP-native distribution, optional fast-routing classification, and dedicated IP infrastructure for reputation isolation.
Inbound replies
Every reply arrives as a structured payload — sender, subject, body, headers, attachments. Injection score + sender reputation always included. Webhook (HMAC-signed) or SDK reactor.
Read more →Per-agent reputation graph
Real engagement signals, not synthetic warmup. Suppression propagates across all your agents, isolated per workspace.
Read more →MCP server
npx @mailsai/mcp-server is live on npm today — drop-in for Claude Code, Cursor, Cline, Continue, Windsurf, OpenAI Agents SDK, Anthropic SDK. Runs client-side, no server to host.
Read more →Classification (opt-in)
Extract intent + entities + urgency on every inbound via our 2-call classifier. Opt-in per agent for +$0.003 per inbound. Skip if your agent runs its own LLM on every reply — you'd be paying twice for the same extraction.
Read more →Dedicated IP add-on
$50/mo per IP on Scale+, available on request. Reputation isolated from the shared pool. A roadmap add-on; current beta sending is on the shared pool.
Read more →Built for agents.
Self-serve in minutes.
The API is live and self-serve. Drop ~6 lines into your agent and ship.
$ npm install @mailsai/sdk