Overview
Virtual Mail gives you programmatic access to your virtual mail — inboxes, mail items, scans, shipments, check deposits, and more. Two interfaces are available depending on how you want to integrate.
MCP server — for AI assistants
Section titled “MCP server — for AI assistants”The MCP server (Model Context Protocol) lets AI assistants like Claude, ChatGPT, and Cursor connect directly to your virtual mailbox. Your assistant can find, read, organize, scan, and shred mail on your behalf, and track shipments and check deposits — all within your own account and permissions.
- Protocol: MCP over stateless streamable HTTP — a single endpoint, no WebSocket upgrade, no session state.
- Auth: OAuth (recommended) or a personal API key.
- Capabilities: read inboxes, account overview, mail items, folders, recipients, and tags. Act on mail: mark read or unread, move, archive, or trash, tag, transfer to another recipient, and request or cancel scans and shredding. Create and delete tags. Shipment and check deposit status is read-only — the MCP server cannot ship mail or deposit checks.
➜ Public MCP Documentation — authentication, capabilities, and how to connect. ➜ Tools Reference — every tool exposed to agents.
Public API — for application integrations
Section titled “Public API — for application integrations”The Public API provides direct HTTP access to virtual-mail resources for custom application integrations, automation scripts, and third-party platforms.
- Protocol: REST over HTTPS — standard JSON request/response.
- Auth: Static API key (generated in Settings → Integrations).
- Capabilities: read inboxes, mail items, scans, media, check deposits, shipping addresses, shipments, users, accounts, and recipients. Act on mail: archive and unarchive, request or cancel shredding, email a scan, get shipping rates, create shipments, and request or cancel check deposits. Create, update, and delete tags, and add or remove them on mail items.
➜ Public API Documentation — authentication, rate limits, paging, and key concepts. ➜ Quickstart — your first request, with ready-to-run example scripts. ➜ Playbooks — step-by-step workflows for common tasks. ➜ API Spec — full OpenAPI 3.1 endpoint reference.
Which one should I use?
Section titled “Which one should I use?”| MCP server | Public API | |
|---|---|---|
| Best for | AI assistants and agentic flows | Custom apps, scripts, and system integrations |
| Protocol | MCP (streamable HTTP) | REST (HTTPS + JSON) |
| Auth | OAuth or API key | Static API key |
| Interface | Tools with structured input/output | HTTP endpoints |
| Spec | Tools Reference | OpenAPI 3.1 Spec |
Both interfaces access the same underlying data and enforce the same per-user authorization. Choose MCP if you’re connecting an AI assistant; choose the Public API if you’re building a traditional integration.