Mailbuttons MCP Server
Enables AI agents to send and receive email with enforced security policies, scoped mailboxes, and human approval for external sending.
README
Mailbuttons MCP server
Governed email for AI agents, over the Model Context Protocol. Give an agent a scoped mailbox where the server enforces the guardrails, not the prompt: sandbox-by-default inboxes, a policy gate on every send, and a tamper-evident audit log. External sending and scope changes are human-approved, never granted by the agent itself.
- Registry:
com.mailbuttons/mcp-server - npm:
@mailbuttons/mcp-server - Platform: mailbuttons.com · agent front door: mbag.ai
Use it
Hosted (recommended) — a streamable-HTTP endpoint served by the platform:
https://mailbuttons.com/api/v1/mcp/rpc
Authorization: Bearer <your Mailbuttons MCP token> # mb_sandbox_... (or mb_prod_...)
Local (stdio) — run the npm package and let your agent launch it:
{
"mcpServers": {
"mailbuttons": {
"command": "npx",
"args": ["-y", "@mailbuttons/mcp-server"],
"env": { "MAILBUTTONS_API_KEY": "mb_sandbox_..." }
}
}
}
Get a scoped token from the dashboard or POST /api/v1/mcp/sandbox-inboxes.
Sandbox tokens can never send externally until a human promotes them.
What the server enforces
- Inbound — a per-mailbox sender policy, fail-closed: mail from strangers bounces by default, so nobody can prompt-inject your agent just by emailing it.
- Outbound — a recipient blocklist the agent can read but not change; a
blocked send returns a normal
{"status":"blocked"}result, not an error. - Caps + audit — per-account send caps and a hash-chained audit log that records refusals too.
- Escalation — sending externally or widening scope is propose-only; a named human approves. The agent cannot approve its own request.
Install as an agent skill
This repo ships a root SKILL.md, so any skills-aware agent can add it:
npx skills add mailbuttons/mcp-server
Docs
- Agent signup (for agents): https://mbag.ai/docs/agent-signup.md
- Trust model: https://mbag.ai/docs/trust-model.md
- OpenAPI: https://mbag.ai/api/openapi.json
- Framework recipes:
references/(Claude Agent SDK, LangChain, plain SDK)
Mailbuttons is a trading name of Code Cutter Limited (UK, no. 08453060). MIT licensed.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.