Comparison — Postfleet vs inbound.new
inbound.new and Postfleet are both hosted inbound-email APIs for agents with webhooks, threading, and an MCP server; they differ in what happens before delivery. inbound.new parses the message and applies the filter rules you configure. Postfleet screens every message for sender auth, hidden content, spam, malware, and prompt injection, then delivers typed JSON.
When an LLM reads the mail and you want the screening enforced by the service rather than configured rule by rule: authentication verdicts on every message, deterministic sanitization, spam and virus gates, an injection scan that fails closed, typed extraction, and approval before send.
When you want an open-source core, a Resend-compatible SDK, and a low flat price for plain parsed email delivered to a webhook, and your own code will handle filtering beyond the rules you write in Inbound Guard.
Pricing model. inbound.new is a flat monthly plan metered on emails, from $4 a month for 5,000 emails as of September 6, 2026, with the Inbound Guard filter included on Pro and above. Its free plan was retired in December 2025, though some marketing pages still mention one. Postfleet is priced per comprehended message, at $0 for 100 a month, $20 for 1,500, and $200 for 25,000 or more, with unmetered sanitization on every plan.
“DIY” means the capability can be assembled on the platform but is not enforced as a built-in service. “Not documented” means the public product documentation reviewed does not substantiate the capability; it does not mean the capability is impossible.
Swipe or scroll to compare →
| Capability | Postfleet | inbound.new |
|---|---|---|
| Comprehension & security | ||
| Prompt-injection screening with fail-closed handling | Yes | Not documented |
| Content sanitization with an audit report | Yes | No |
| JSON Schema extraction with confidence scores | Yes | Not documented |
| Automatic message classification | Yes | DIY |
| Blocking wait for a matching email | Yes | Not documented |
| PDF text included in message processing | No | Not documented |
| Inbound SPF / DKIM / DMARC verdicts surfaced | Yes | Not documented |
| Layered spam gate with quarantine and recovery | Yes | DIY |
| Virus gate with fail-closed delivery | Yes | Not documented |
| Async attachment byte scanning before download | No | Not documented |
| Sending controls & delivery reliability | ||
| Server-enforced human approval before send | Yes | No |
| Automatic webhook retries | Yes | Yes |
| Dead-letter queue with self-serve redrive | Yes | DIY |
| Inbox essentials | ||
| MCP server (hosted + local) | Yes | Yes |
| REST API | Yes | Yes |
| Webhooks | Yes | Yes |
| Send + receive | Yes | Yes |
| Custom domains (DKIM / SPF / DMARC) | Yes | Yes |
| Threading | Yes | Yes |
| Mailbox-scoped API keys with read / send flags | Yes | Not documented |
| Send allowlists / blocklists | Yes | Not documented |
| Idempotent message sends | Yes | Not documented |
| Where inbound.new is broader today | ||
| Open-source core under the MIT license | No | Yes |
| Client SDK for the REST API (TypeScript, on npm) | No | Yes |
| Resend-compatible SDK surface for migration | No | Yes |
| Forward inbound mail to an address or a group of up to 50 | No | Yes |
| Natural-language generation of filter rules | No | Yes |
Not in the documented webhook payload. Domain setup reports the DNS records to publish and a verified status, but no per-message authentication verdict field appears in the published schema. Postfleet attaches SPF, DKIM, and DMARC verdicts to every inbound message before the agent reads it.
It ships Inbound Guard, a rule engine you configure with explicit patterns or natural-language prompts to block, flag, or route mail, included on Pro plans and above. It is a spam filter you write rules for, not an always-on gate, and virus scanning or prompt-injection screening is not documented. Postfleet runs those gates on every message by default.
Yes. A hosted server runs at mcp.inbound.new and a self-hostable package supports HTTP and stdio transports, with documented setup for Claude Desktop, Cursor, and OpenCode. Its tools cover domains, endpoints, sending and retrieving email, and threads. Postfleet also ships hosted and local MCP, and its read tool returns the screened message rather than the raw body.
No. The company retired its free plan in December 2025, citing costs as a bootstrapped business, and the cheapest plan is now $4 a month for 5,000 emails. Some marketing pages still advertise free emails; the pricing page is the current source. Postfleet keeps a free tier with no card required.
The core application is published under the MIT license on GitHub, with a local development mode for testing webhooks without AWS. A production self-hosting guide was not found, so treat the repository as the source of the hosted product rather than a documented deployment path. Postfleet does not publish its core as open source.
The same table and the same legend, against the other ways teams give an agent an inbox.
Last verified September 6, 2026. We compared documented product behavior, not what could be built with additional infrastructure. inbound.new links point to its primary documentation; Postfleet links point to the product pages that describe the corresponding controls.