Cloudflare Email Service

Original Article ↗ Hacker News Discussion ↗

Cloudflare has launched a new transactional email service, tightly integrated with its Workers platform and marketed as a way for “AI agents” to send and receive email. Commenters see value in an SES-style alternative—especially given SES onboarding pain and higher prices from providers like SendGrid or Resend—but are sharply divided over Cloudflare’s pricing, lack of SMTP at launch, and whether its abuse handling is strong enough to prevent the service from becoming a large-scale spam vector. Many are also skeptical of the heavy “agent” framing, arguing that it adds hype and potential for AI-generated spam to what is essentially a conventional email API.

Scope and Positioning of the Service

  • Many see this as Cloudflare’s SES-style transactional email API, tightly integrated with Workers and the broader Cloudflare “AWS competitor” stack.
  • Marketing emphasis on “agents” is widely viewed as hype; commenters argue it’s just an email API that can be used by anything, including LLM agents.
  • Some are confused by the framing and expected a full consumer email platform (Gmail/365 competitor), which this is not.
  • Current sending appears to be via Workers/REST; lack of SMTP support is a blocker for some workflows, though docs suggest SMTP is planned.

Pricing and Comparisons

  • Advertised pricing: ~$0.35 per 1,000 outbound emails, unlimited inbound, plus daily limits tied to account standing, and only available on Workers Paid.
  • Compared to AWS SES (~$0.10 per 1,000), Cloudflare is ~3x more expensive, but cheaper than many SaaS senders (SendGrid, Postmark, Resend).
  • Some appreciate the pay-per-use model versus fixed monthly tiers; others note SES’s attachment pricing can narrow the gap.
  • Several say they’d gladly switch from current providers (especially SendGrid/Resend) if deliverability and tooling are good.

Spam, Abuse, and Reputation Concerns

  • Strong fear this becomes “just another spam source,” especially for AI/agent-generated cold outreach and “spin up warm-up inboxes.”
  • Some note you can’t practically block “all of Cloudflare,” which is seen as part of the appeal for senders and a problem for recipients.
  • Cloudflare’s historic anti-abuse posture is criticized (referencing Spamhaus stats and piracy/La Liga blocking episodes), raising doubts they’ll police email abuse aggressively.
  • Others argue that to maintain deliverability Cloudflare must and will fight spam, with reserved IPs and monitoring, similar to SES.
  • Email professionals highlight that at scale most effort is abuse mitigation; opinions differ on how “rocket science” deliverability really is.

Impact on Ecosystem and Startups

  • Several see this as bad news for small “agent email” startups with thin moats; Cloudflare can absorb the feature as a cheap platform primitive.
  • Broader reflections on email as a “tragedy of the commons”: near-zero-cost sending invites abuse, pushing power to big inbox providers.
  • Ideas like proof-of-work, payments/tips, and better reputation systems are discussed but seen as hard to deploy in practice.

Reception and Use Cases

  • Enthusiasts want an alternative to SES’s opaque approval process and to “enshittified” transactional providers.
  • Some experiment with email as an interface to LLM agents (e.g., threaded workflows, bookkeeping, incident triage) and are excited.
  • Others reject agents-in-email outright, see this as contributing to a “dead internet” of bots emailing bots, and vow to aggressively block such mail.