radmail-ai/radmail-mcp

Overview

Email OS for AI agents over MCP. Provides two-axis triage, a Right Now lane, commitment follow-through, reviewable drafts, and a machine-verifiable BEC hard-stop that forces human-only handling for money/banking changes or first contact. Includes a zero-auth sandbox for testing.

Best for

Best for
AI agent developers building email automation with strong security guardrails.

Use cases

• Automate email triage and prioritization
• Track and follow through on commitments from emails
• Draft and review emails with agent assistance

Notes

Email OS for AI agents over MCP. Provides two-axis triage, a Right Now lane, commitment follow-through, reviewable drafts, and a machine-verifiable BEC hard-stop that forces human-only handling for money/banking changes or first contact. Includes a zero-auth sandbox for testing.

0 stars on GitHub. Last updated 2026-09-02. Licensed MIT.

Use cases

• Automate email triage and prioritization
• Track and follow through on commitments from emails
• Draft and review emails with agent assistance

Pros

• Machine-verifiable BEC protection reduces fraud risk
• Zero-auth sandbox enables safe testing
• Reviewable drafts give human oversight

Cons

• Zero-auth sandbox may not cover real-world authentication flows
• Requires MCP integration, limiting non-MCP environments
• BEC hard-stop adds friction for legitimate urgent requests

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.