mcpField
B
BumbleBox

BumbleBox

Remote
bumblebox.aiCommunication0Updated 1d ago

About BumbleBox

BumbleBox is a remote email-intelligence MCP server for people managing multiple businesses. It searches connected mailboxes and supports triage, drafts, campaigns, and controlled sending.

Key features

  1. 01Search and read connected mailboxes
  2. 02Find emails awaiting replies
  3. 03Access contacts and calendar events
  4. 04Draft replies and campaigns
  5. 05Organize email with reversible actions
  6. 06Send only with configured approval

Use cases

  1. 01Find a message across several mailboxes
  2. 02Review emails that need a response
  3. 03Prepare a reply for approval
  4. 04Organize mail without permanent deletion

MCP support

Exposes email search, triage, drafting, campaign, and mailbox-management tools to MCP clients. OAuth 2.1 controls access, and sending remains subject to approval settings.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://api.bumblebox.ai/api/mcp
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

Tools

Details

From the project README

BumbleBox MCP server

Email intelligence for people running several businesses, inside your AI assistant. BumbleBox triages every connected mailbox (Gmail, Outlook, any IMAP); this remote MCP server lets Claude, ChatGPT, Cursor, VS Code or any MCP client work with it.

  • Endpoint: https://api.bumblebox.ai/api/mcp (streamable HTTP)
  • Auth: OAuth 2.1 with dynamic client registration and PKCE. You approve once on bumblebox.ai and see exactly which permissions the assistant gets.
  • Plans: Pro or Business (pricing).
  • Setup for every client: https://bumblebox.ai/integrations

Install

Claude Code (plugin, with the /bumblebox:to-answer and /bumblebox:tidy commands):

/plugin marketplace add BumbleBox/bumblebox-mcp
/plugin install bumblebox@bumblebox

Or just the server: claude mcp add --transport http bumblebox https://api.bumblebox.ai/api/mcp

Claude (claude.ai, desktop): Settings → Connectors → Add custom connector → https://api.bumblebox.ai/api/mcp.

ChatGPT: Settings → Apps & Connectors → create a connector (developer mode), paste the server URL, authentication OAuth.

Cursor (~/.cursor/mcp.json):

{ "mcpServers": { "bumblebox": { "url": "https://api.bumblebox.ai/api/mcp" } } }

VS Code (.vscode/mcp.json):

{ "servers": { "bumblebox": { "type": "http", "url": "https://api.bumblebox.ai/api/mcp" } } }

Tools

ToolWhat it does
search_emails, get_emailSearch and read every connected mailbox
list_to_answerEmails waiting for your reply
search_contacts, list_folders, list_calendar_eventsContacts, folders, calendar
draft_replyA reply draft that waits in BumbleBox for your click on Send
send_replyA reply to the original sender, at the level you set in BumbleBox
archive_email, move_email, trash_email, restore_to_inboxTidy on request, always undoable
list_missions, list_campaign_drafts, list_replies_to_validateYour campaigns
create_campaign_draft, update_campaign_draftCampaign drafts, sent only from BumbleBox

What it never does

BumbleBox sends nothing without your click — Send, Schedule, or a standing permission you set yourself in BumbleBox for a precise scope (existing participants of a thread you opened, a message with no commitment), capped, logged and cancellable. Anything else waits for your validation, and an assistant can never change that permission.

There is no permanent deletion: trash_email moves only certified noise to the Trash and turns anything else into a one-tap proposal in BumbleBox; everything is undone with restore_to_inbox.

Links

This repository holds the public metadata (registry server.json, Claude Code plugin). The server itself is operated by BumbleBox. Licence: MIT.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.