Postfleet

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 — your agent gets an explicit status (skipped_injection_risk) 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 — retry with the same client_id and the email goes out at most once — 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.