Thunderbird
Microsoft Outlook
Mailbird
eM Client
Postbox
Mailspring
ProtonMail
Apple Mail
Postfleet
Nylas Mail
Evolution
Mailbird
Mailpile
Polymail
Robotomail
Postfleet is inbound email infrastructure for AI agents, built on the assumption that any message can be an attack.
Every message runs the same pipeline before your agent sees it: SPF/DKIM/DMARC checks, HTML-to-text cleaning, removal of hidden elements, HTML comments and zero-width characters, malware and spam screening, then a prompt-injection scan. What comes out is typed JSON matching a schema you define.
It fails closed. Malware is quarantined, spam is suppressed, and a message flagged for injection risk never reaches extraction โ your agent gets an explicit status (skippedinjectionrisk) instead of quietly poisoned context. The pre-sanitization body is never sent to agent-facing tools.
Connect over MCP with 14 tools (Claude, Claude Code, Cursor, Windsurf, Cline, VS Code) or use the REST API. Sends are idempotent โ retry with the same client_id and the email goes out at most once โ with optional human approval before anything irreversible leaves.
We publish the red-team scoreboard, including what got through and one false flag we caught ourselves: 99.1% of 113 attacks caught, most written by independent models and a public benchmark rather than by us. Screened against known patterns, never claimed immune.
Free: 100 messages, 3 mailboxes. Sanitization and injection screening are never metered.
Thunderbird PostfleetNo Postfleet videos yet. You could help us improve this page by suggesting one.
Postfleet's answer:
It treats an agent's inbox as a trust boundary, not a pipe. Every inbound email is screened for prompt injection, stripped of hidden text, and turned into structured JSON before the agent ever reads it. The unusual part: I red-team the filter with rival models and publish the whole scoreboard, catches and misses both. Screened, never immune.
Postfleet's answer:
Most email APIs for agents hand you the raw body and leave defense to your app layer. Postfleet screens first and fails closed, so flagged mail reaches the agent as inert data instead of trusted content. It's backed by an injection guarantee and a public red-team scoreboard, so the security claim comes with receipts rather than a marketing number.
Postfleet's answer:
Developers building AI agents that receive email: indie hackers, agent-framework builders, and small platform teams wiring an agent to an inbox. People who've read at least one prompt-injection horror story and would rather see published results than take a claim on faith.
Postfleet's answer:
I've spent 18 years building software, the last 6 as an architect drawing trust boundaries between systems that shouldn't blindly trust each other. When I watched people give AI agents email inboxes, I saw the boundary everyone skipped: anyone who can email an agent can drop instructions into its context. Postfleet is that missing boundary.
Postfleet's answer:
Next.js, TypeScript, and Supabase (Postgres), deployed on Vercel. An MCP server so an agent can sign itself up. Anthropic models run the injection scan and schema extraction, Resend/SES handles inbound mail, and ClamAV scans attachments.
Microsoft Outlook - Organize your world. Outlookโs email and calendar tools help you communicate, stay on top of what matters, and get things done.
Nylas Mail - The Nylas Cloud API powers your application with email, calendar & contacts features. Built-in features for better email, calendar, and contact management.
Mailbird - Mailbird is the best email client for Windows 7, 8 and 10
Evolution - Apps/Evolution - GNOME Wiki!
eM Client - eM Client is a fully-featured email client for Windows and macOS with a clean and easy-to-use interface. eM Client also offers features for calendars, tasks, contacts, notes, and chat.
Postbox - The Power Email App for busy professionals, like you.