This connects Claude to Thrd's agent-first email infrastructure through tools for polling inbound events, inspecting threads, and sending or replying with policy-enforced safety controls. It exposes the core Thrd API operations: provision isolated agent inboxes, consume normalized email events via long-poll, ACK processed batches, and trigger outbound messages that pass through deterministic prompt shields before sending. Reach for this when you're building autonomous agents that need to handle real email workflows without exposing your personal inbox or wrestling with raw SMTP chaos. The MCP wraps the same REST primitives documented in their OpenAPI spec, so you get thread-locked replies, proof of receipt hashes, and tiered trust gates without manually wiring HTTP clients.
claude mcp add --transport stdio sergiorico1-thrd -- npx -y thrd-mcp