# Mailbuttons

> Governed email for AI agents, hosted in the EU/UK. Policy screens every inbound message before an agent sees it; outbound sends pass capability checks and tier caps; every tool call lands in a tamper-evident audit chain. Built for teams that need governance, data residency, and evidence.

## API

- [OpenAPI specification](https://mailbuttons.com/api/openapi.json): the /api/v1 REST surface — mailboxes, send, threads and messages, policy, webhooks, memory files, JMAP proxy, and the OTP signup flow

## MCP server

- Endpoint (Streamable HTTP, JSON-RPC): https://mailbuttons.com/api/v1/mcp/rpc, header `Authorization: Bearer mb_sandbox_...`
- 12 governed tools: list/read messages and threads, attachment text, send or draft email, held drafts, one-time-code extraction, sender proposals, promotion requests, audit tail and export
- Get a free token: POST https://mailbuttons.com/api/auth/otp/request {"email"}, then POST https://mailbuttons.com/api/auth/otp/cli-exchange {"email","code"} returns an mbk_ API key; POST https://mailbuttons.com/api/v1/mcp/sandbox-inboxes with `Authorization: Bearer mbk_...` returns a sandbox inbox and its sandbox_token
- Claude Code: `claude mcp add --transport http mailbuttons https://mailbuttons.com/api/v1/mcp/rpc --header "Authorization: Bearer mb_sandbox_..."`
- Local stdio: `npx -y @mailbuttons/mcp-server` with MAILBUTTONS_API_KEY set to the token

## Docs (raw markdown)

- [Webhooks](https://mailbuttons.com/docs/webhooks.md)
- [Mail policy](https://mailbuttons.com/docs/policy.md)
- [Audit log](https://mailbuttons.com/docs/audit-log.md)

## Site

- [Why Mailbuttons](https://mailbuttons.com/why-mailbuttons/)
- [Trust & security](https://mailbuttons.com/trust/)
- [Pricing](https://mailbuttons.com/pricing/)
- [Developers](https://mailbuttons.com/developers/)
- [Agent skill](https://mailbuttons.com/.well-known/skills/index.json)