# Envolt Mail

> Mail built for AI agents and the humans that build them. Envolt Mail is an email platform being built so that every artificial intelligence (AI) agent has its own clearly labeled, verifiably sealed mailbox, and an accountable person decides what it sends. In development. The core inbound pipeline runs in staging on synthetic test mail; nothing is sent and no real mailbox is connected.

Every page is also available as Markdown: add `.md` to its path (`https://envoltmail.com/index.md` for the home page) or send `Accept: text/markdown`. The whole site in one file: https://envoltmail.com/llms-full.txt

## Core URLs
- [Home](https://envoltmail.com/index.md): Envolt Mail is being built so every AI agent has a clearly labeled, verifiably sealed mailbox, and an accountable person decides what it sends.
- [Status](https://envoltmail.com/status.md): What Envolt Mail runs today, what is only designed, the latest release evidence, and what comes next. Built from one status file.
- [How it works](https://envoltmail.com/how-it-works.md): From inbound mail to a sealed, labeled reply: each step of Envolt Mail, and which steps already run in staging.
- [Security](https://envoltmail.com/security.md): How Envolt Mail isolates customers, encrypts mail, resists prompt injection and keeps people accountable, with what runs today.
- [For builders](https://envoltmail.com/agents.md): What Envolt Mail is designed to give AI agents and the teams building them: identity, safe inbound mail, a guarded outbox and earned autonomy.
- [Docs](https://envoltmail.com/docs.md): Envolt Mail documentation: how recipients recognize and verify an agent, the builder preview, a glossary and answers to common questions.
- [Recipient guide](https://envoltmail.com/docs/recipients.md): How to tell a message was written by an AI agent on Envolt Mail, how to verify it, and how to reach a person instead.
- [Builder preview](https://envoltmail.com/docs/builders.md): Preview for builders: the Envolt-Agent-Seal header, its canonical signing string, the public key document, and the planned API, webhooks and MCP tools.
- [Glossary](https://envoltmail.com/docs/glossary.md): Plain-language definitions of the terms Envolt Mail uses: Airlock, Agent Seal, Seal policy, trust tiers, receipts, ledger anchors and more.
- [FAQ](https://envoltmail.com/faq.md): Answers about Envolt Mail: availability, your mail server, whether agents can send on their own, AI models, pricing and security.
- [Changelog](https://envoltmail.com/changelog.md): What changed in Envolt Mail, in plain language, newest first. Service entries point at the release receipt that proves them.
- [About](https://envoltmail.com/about.md): Why Envolt Mail exists, who builds it, and how we build it: AI agents need email, and someone has to stay accountable for every message.

## What this project does
Envolt Mail is designed so that every agent message is labeled as coming from an agent and carries a verifiable Agent Seal. Agents propose; a policy check and an accountable person decide what is sent. Content inside an email is treated as data, never as instructions. Today the inbound pipeline (quarantine, hidden-text removal, Agent Seal verification, a no-tools reader and a tamper-evident ledger) runs in staging on synthetic test mail; nothing is sent and no real mailbox is connected.

## Studio
A VaultSpark Studios LLC project. https://vaultsparkstudios.com

## Optional
- Contact and early access: https://envoltmail.com/contact (hello@envoltmail.com)
- Changelog feed: https://envoltmail.com/changelog.xml
- Privacy: https://envoltmail.com/privacy
- Terms: https://envoltmail.com/terms