Last updated 08-21-2026
Category:
Reviews:
Join thousands of AI enthusiasts in the World of AI!
Mails.ai
Mails.ai gives software agents a real email inbox through a REST API and npm SDK. Create an agent address, send transactional messages, receive replies, and hook replies into your code with structured events that include injection scores and sender reputation.
Generic email APIs leave cold outreach enforcement to an acceptable-use policy your agent can accidentally violate. Mails.ai refuses cold outreach inside the send call with a 422 response, so a bad prompt cannot burn your domain reputation. That API-level block is the clearest reason to pick it over a general SMTP relay when agents are doing the sending.
Developers building support bots, booking assistants, and workflow agents use it with TypeScript, Python, or the MCP server in Claude Code and Cursor. Pricing starts free with no card, then monthly Pro and Scale tiers while metered per-event billing rolls out.
Creates a working agent inbox with about six lines of SDK code
Returns 422 cold_email_prohibited instead of sending unsolicited outreach
Scores every inbound message for prompt injection before your agent reads it
Optional classification adds intent, entities, and urgency for $0.003 per inbound on metered pricing
Ships @mailsai/sdk and @mailsai/mcp-server on npm for TypeScript, Python, and MCP workflows
Monthly tiers today at Free $0, Pro $20, and Scale $99 while metered billing launches
Blocks cold outreach inside the send API instead of relying on policy text alone.
Every inbound arrives with injection_score and sender_reputation fields your code can trust.
npm SDK and MCP server make it easy to wire email into Claude Code or Cursor agents.
Metered per-event pricing is marked coming soon, so long-term cost math is still in flux.
Classification covers 12 canonical intents, which may not fit niche domain-specific routing.
English-first product copy; multilingual agent mail is not highlighted on the site.
What does Mails.ai do for AI agents?
Mails.ai provisions a real email address for each agent through its API. Agents can send transactional mail, receive replies, and consume structured reply events with injection scores, sender reputation, and optional intent parsing.
Does Mails.ai block cold email?
Yes. Mails.ai rejects cold outreach during the send API call with a 422 cold_email_prohibited response. The refusal happens before delivery, so agents cannot accidentally send unsolicited pitches through the service.
Is there a free plan on Mails.ai?
Yes. Mails.ai lists a Free plan at $0 with self-serve signup. Paid monthly Pro and Scale plans are listed at $20 and $99 while metered per-event pricing rolls out.
How does Mails.ai scan inbound mail?
Mails.ai runs a six-category prompt-injection scanner on every inbound message. Each reply event includes an injection_score and sender_reputation field, and high-risk messages can be quarantined before your agent acts on them.
Can I use Mails.ai with MCP tools?
Yes. Mails.ai publishes @mailsai/mcp-server on npm alongside @mailsai/sdk, and the site lists integrations with Claude Code, Cursor, Cline, Windsurf, OpenAI Agents, and LangGraph.
What does Mails.ai classification add?
Mails.ai classification is opt-in per agent and extracts intent, entities, and urgency from inbound replies. On metered pricing it adds $0.003 per classified inbound on top of the base $0.002 inbound scan.