Verified, deduplicated ingest
Postfleet verifies signed webhooks, deduplicates provider retries, and uses idempotent ingest so repeated delivery events do not create duplicate billing or duplicate processing.
Postfleet
Postfleet is email infrastructure for AI agents that cleans, screens, and structures inbound mail before an agent reads it. Supports hosted and local MCP access, plus REST integration, with deterministic sanitization on every message.
What Postfleet is
Postfleet is email infrastructure for AI agents. It sits between inbound mail and the agent that consumes it, turning raw messages into cleaned, screened, structured data before the agent reads a word.
The product is designed around the idea that an inbox is a trust boundary. It verifies and deduplicates inbound delivery, authenticates messages, strips noise and hidden content, scans for malware and spam, checks for prompt injection, and only then extracts clean mail into typed JSON with a confidence score.
Core capabilities
Message authentication checks
Inbound mail is checked for SPF, DKIM, and DMARC verdicts, and those authentication results are surfaced with the message.
Deterministic sanitization pipeline
HTML is converted to text and deterministic cleanup removes tracking pixels, styling noise, quoted reply chains, hidden text, zero-width characters, and homoglyph runs.
Attachment and malware gating
SES malware verdicts gate delivery, attachments are scanned asynchronously at the byte level, and failed scans are treated as unsafe rather than clean.
Fail-closed message gating
Spam is quarantined by default and can be reclassified and redelivered by an operator, while injection-risk messages are retained without extraction.
Structured extraction to JSON
A model then extracts clean mail to typed JSON with a confidence score, and the dashboard shows raw versus comprehended content side by side.
Practical use cases
Agent inbox processing
Use Postfleet when an agent needs to read inbound email but should only receive cleaned, structured content after authentication, sanitization, and prompt-injection screening.
Schema-based email extraction
Use the platform for workflows that need incoming messages mapped into a predefined JSON schema, such as extracting fields from receipts, replies, or request emails.
Inbound email safety filtering
Use the safety pipeline when you want hidden text, zero-width characters, HTML noise, spam, and malware filtered before an agent sees the message.
MCP-based agent setup
Use the hosted MCP endpoint or the local npx package when you want an agent to provision and manage mail flow through an MCP client without building a separate inbox layer.
Message review and auditing
Use the dashboard when operators need to compare raw and comprehended mail, inspect verdicts, and review sanitization records for individual messages.
Pros and Cons
Pros
- Covers the full path from ingest to structured output, including authentication checks, sanitization, spam handling, malware gating, and injection screening.
- Safety processing is deterministic and runs on every message on every tier, including Free.
- Offers both hosted MCP and local npx access, plus REST for teams that prefer to integrate without MCP.
- Plans are tiered for different stages of use, from a free test tier to Pro and Scale.
- Includes a 30-day money-back guarantee on Pro and an injection guarantee tied to the fail-closed gate.
Cons
- Comprehension is metered, so extraction pauses for the rest of the month once the plan limit is reached.
- The site does not publish a full list of third-party integrations beyond MCP and REST.
- The public pages do not show detailed output schema examples or a step-by-step onboarding flow beyond the quickstart snippets.
FAQ
What does Postfleet do?
Postfleet provides email infrastructure for AI agents. It ingests inbound mail, cleans and screens it, and then extracts clean messages to a JSON schema before your agent reads them.
How do you connect an agent to Postfleet?
The hosted MCP endpoint and the local npx package both use the same pf_ key and the same plan limits. You can connect through MCP or, if you prefer, drive it over REST.
What is a comprehended message?
A comprehended message is an inbound message that passes through the full LLM stage: prompt-injection scan, then extraction to your JSON schema with a classification and confidence score.
Is safety metered?
No. Sanitization is deterministic and runs on every message on every tier, including Free. The metered part is comprehension, which is the LLM extraction stage.
Can you cancel or change plans?
Yes. The pricing page says you can cancel anytime from the dashboard, and Pro includes a 30-day money-back guarantee.
Quick Facts
- Category
- Developer Tool
- Primary use
- Email infrastructure for AI agents
- Access methods
- Hosted MCP, local npx MCP, REST
- Pricing page
- Free, Pro, and Scale plans are listed
- Safety model
- Deterministic sanitization on every message; prompt-injection screening before extraction
- Website
- postfleet.ai
Postfleet Alternatives
CreateOS Sandbox
CreateOS Sandbox is an isolated compute environment for running code and agent workloads in Firecracker micro-VMs with private networking and SDK, CLI, or MCP control.
AakarDev AI
AakarDev AI helps teams manage AI provider access, project setup, logs, and analytics in one dashboard. BYOK support included.
AgentMail
AgentMail is an email inbox API for AI agents to create, send, receive, and search messages via REST APIs and SDKs. Built for threaded workflows, verification, support, scheduling, and approvals.
Arduino VENTUNO Q
Arduino VENTUNO Q is an edge AI computer for AI and robotics, combining AI inference and deterministic control on one board with Arduino App Lab support.
Devin
Devin is an AI coding agent and software engineer for planning and executing complex software tasks, with desktop, cloud, JetBrains, and CLI access.
ByteAsk
ByteAsk is a terminal-first AI coding agent for C and C++ that edits repos and verifies changes with compilers, debuggers, sanitizers, and tests.