Cloudflare Email MCP Server
Manages Cloudflare KV namespaces and automates email processing from ProtonMail Bridge to structured KV storage with folder mapping.
README
Cloudflare Email MCP Server
A comprehensive Model Context Protocol server for Cloudflare KV operations and automated email processing from ProtonMail Bridge to Cloudflare KV namespaces.
Features
KV Operations
- Key Management: List, count, get, put, delete keys in any KV namespace
- Pagination: Handle large datasets with cursor-based pagination
- Bulk Operations: Retrieve multiple keys efficiently
Email Processing
- Automated Transfer: Move emails from specific IMAP folders to corresponding KV namespaces
- UUID v5 Body IDs: Deterministic body identification with separate storage
- Folder Mapping: Pre-configured mappings for court, government, and legal correspondence
- Status Tracking: Pending/processed status for workflow management
Architecture
Two-Tier Storage
Court KV Namespace:
├── Key: 2026.02.09_casework_ico_org_uk_10-30-45
└── Value: {
"from": "casework@ico.org.uk",
"to": "rose@mobicycle.ee",
"subject": "Your complaint reference IC-...",
"body": "f47ac10b-58cc-4372-a567-0e02b2c3d479",
"namespace": "ICO Complaints",
"status": "pending"
}
Body KV Namespace (by year):
├── Key: f47ac10b-58cc-4372-a567-0e02b2c3d479
└── Value: { "text": "Dear Ms Liu..." }
Folder Mappings
Automatically processes emails from:
- Courts: Appeal, Chancery, Admin Court, King's Bench, Supreme, etc.
- Government: GLD, US State Dept, Estonian Government
- Complaints: HMCTS, ICO, PHSO, Bar Standards, Parliament
- Parties: Liu Litigation, HK Law, Lessel Property, etc.
Setup
Prerequisites
- Bun runtime
- ProtonMail Bridge running locally
- Cloudflare account with KV access
- Wrangler CLI configured
Authentication
The server uses Wrangler OAuth tokens automatically:
wrangler auth login
Installation
bun install
bun run build
bun start
Usage
MCP Tools
KV Operations
// Count keys in a namespace
kv_keys_count({
namespace_id: "your-namespace-id",
prefix: "email:"
})
// List keys with pagination
kv_keys_list({
namespace_id: "your-namespace-id",
prefix: "2026.02",
limit: 100,
cursor: "next-page-cursor"
})
// Get/set individual values
kv_key_get({ namespace_id: "...", key: "..." })
kv_key_put({ namespace_id: "...", key: "...", value: "..." })
Email Processing
// Transfer emails from approved folders
email_transfer_folders({
folders: ["INBOX/Courts/Supreme"],
since: "2026-02-01T00:00:00Z",
dry_run: true
})
// Check processing stats
email_folder_stats({
folder_name: "INBOX/Complaints/ICO"
})
// List all folders and their approval status
email_list_folders()
Configuration
Approved Folders
Only pre-configured folders automatically sync to KV:
- Legal correspondence folders
- Court communications
- Government agencies
- Complaint systems
KV Namespaces
- Court namespaces: One per jurisdiction
- Body storage: Separated by year (
email-bodies-2026) - Account mapping: Bridge gRPC account information
Development
Project Structure
src/
├── index.ts # Main MCP server with stdio transport
├── cloudflare-api.ts # KV operations via Cloudflare API
└── email-processor.ts # IMAP to KV sync logic
dist/ # Built JavaScript
README.md # This file
package.json # Dependencies and scripts
Building
bun run build # TypeScript compilation
bun run dev # Watch mode
Testing
# Dry run email processing
email_transfer_folders({ dry_run: true })
# Test KV operations
kv_keys_count({ namespace_id: "test-namespace" })
Integration
Claude Code MCP Configuration
Add to Claude Code MCP settings:
{
"mcpServers": {
"cloudflare-email": {
"command": "bun",
"args": ["start"],
"cwd": "/path/to/this/directory"
}
}
}
Environment Variables
# Optional - detected automatically from wrangler
CLOUDFLARE_API_TOKEN=your-token
CLOUDFLARE_ACCOUNT_ID=your-account-id
Automation
Bridge Integration
Works with ProtonMail Bridge gRPC for account mapping:
- Automatic token capture on Bridge restart
- Account structure pushed to KV
- Prevents email double-counting
Workflow
- Bridge receives emails in configured folders
- MCP server syncs new emails to KV namespaces
- AI triage system processes pending emails
- Status updates track workflow progress
License
Internal MobiCycle OÜ project - Not for public distribution
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。