Twilio MCP Server
Enables AI-powered SMS messaging through Twilio with automatic conversation threading, message status tracking, and webhook support for receiving inbound messages.
README
Twilio MCP Server
A Model Context Protocol (MCP) server for Twilio that enables AI-powered SMS messaging with intelligent conversation management.
Features
- Send SMS - Send text messages via Twilio with automatic conversation threading
- Receive SMS - Process inbound messages via webhooks with automatic storage
- Conversation Management - Track and manage multi-turn SMS conversations
- Message Status - Check delivery status of sent messages
- Thread Association - Automatically link messages to conversation threads
- MMS Support - Send and receive multimedia messages (optional)
Installation
# Clone the repository
git clone <repository-url>
cd twilio-mcp
# Install dependencies
npm install
# Build the project
npm run build
Configuration
- Copy the environment template:
cp .env.example .env
- Fill in your Twilio credentials in
.env:
# Get these from https://console.twilio.com
TWILIO_ACCOUNT_SID=ACxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
TWILIO_AUTH_TOKEN=your_auth_token_here
TWILIO_PHONE_NUMBER=+1234567890
# Webhook configuration
WEBHOOK_PORT=3000
WEBHOOK_BASE_URL=https://your-domain.com
# Database path
DATABASE_PATH=./data/twilio.db
Setting Up Webhooks
For the server to receive inbound SMS, you need to configure Twilio webhooks:
Development (Using ngrok)
-
Install ngrok: https://ngrok.com/download
-
Start the webhook server:
npm run dev
- In another terminal, start ngrok:
ngrok http 3000
-
Copy the HTTPS URL from ngrok (e.g.,
https://abc123.ngrok.io) -
Configure Twilio webhook:
- Go to https://console.twilio.com/us1/develop/phone-numbers/manage/incoming
- Select your phone number
- Under "Messaging Configuration":
- Webhook URL:
https://abc123.ngrok.io/webhooks/twilio/sms - HTTP Method: POST
- Webhook URL:
- Under "Status Callback URL":
- URL:
https://abc123.ngrok.io/webhooks/twilio/status
- URL:
- Save
Production
Deploy the webhook server to a production environment with HTTPS and configure the Twilio webhook URLs accordingly.
Usage with Claude Code
Add this server to your Claude Code MCP configuration:
{
"mcpServers": {
"twilio": {
"command": "node",
"args": ["/path/to/twilio-mcp/dist/index.js"],
"env": {
"TWILIO_ACCOUNT_SID": "ACxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
"TWILIO_AUTH_TOKEN": "your_auth_token",
"TWILIO_PHONE_NUMBER": "+1234567890",
"WEBHOOK_PORT": "3000",
"WEBHOOK_BASE_URL": "https://your-webhook-url.com",
"DATABASE_PATH": "./data/twilio.db"
}
}
}
}
MCP Tools
send_sms
Send an SMS message via Twilio.
Parameters:
to(required): Recipient phone number in E.164 format (+1234567890)message(required): SMS content (1-1600 characters)from(optional): Sender phone number (uses default if not provided)conversationId(optional): UUID of existing conversation
Example:
{
"to": "+1234567890",
"message": "Hello from Twilio MCP!"
}
Response:
{
"messageSid": "SMxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
"status": "queued",
"to": "+1234567890",
"from": "+1987654321",
"conversationId": "550e8400-e29b-41d4-a716-446655440000",
"timestamp": "2025-01-15T10:30:00.000Z"
}
get_inbound_messages
Query received SMS/MMS messages from storage.
Parameters:
from(optional): Filter by sender phone numberto(optional): Filter by recipient phone numberconversationId(optional): Filter by conversation UUIDsince(optional): ISO 8601 timestamp to filter messages after this datelimit(optional): Maximum messages to return (default: 50)
Example:
{
"from": "+1234567890",
"limit": 10
}
create_conversation
Initialize a new conversation thread.
Parameters:
participants(required): Array of phone numbers in E.164 format (min 2)metadata(optional): Custom metadata object
Example:
{
"participants": ["+1234567890", "+1987654321"],
"metadata": {
"campaign": "support",
"priority": "high"
}
}
get_conversation_thread
Retrieve full conversation history with all messages.
Parameters:
conversationId(required): UUID of the conversationincludeContext(optional): Include AI context summary (default: false)
Example:
{
"conversationId": "550e8400-e29b-41d4-a716-446655440000",
"includeContext": true
}
get_message_status
Check the delivery status of a sent message.
Parameters:
messageSid(required): Twilio message SID
Example:
{
"messageSid": "SMxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
}
Response:
{
"messageSid": "SMxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
"status": "delivered",
"timestamp": "2025-01-15T10:31:00.000Z",
"to": "+1234567890",
"from": "+1987654321"
}
Conversation Threading
The server automatically creates and manages conversation threads:
-
Outbound Messages: When sending an SMS:
- If
conversationIdis provided, the message is linked to that conversation - If not provided, the server finds an existing conversation by participant pair
- If no conversation exists and
AUTO_CREATE_CONVERSATIONS=true, a new one is created
- If
-
Inbound Messages: When receiving an SMS via webhook:
- The server matches by participant pair (from/to numbers)
- If no conversation exists and
AUTO_CREATE_CONVERSATIONS=true, a new one is created - Messages are automatically stored and linked to the conversation
-
Participant Matching: Conversations are identified by normalized participant pairs:
[+1234567890, +1987654321]matches the same conversation regardless of direction
Architecture
twilio-mcp/
├── src/
│ ├── index.ts # MCP server entry point
│ ├── server.ts # MCP server with tool registry
│ ├── webhook-server.ts # Express server for webhooks
│ ├── config/
│ │ └── env.ts # Environment configuration
│ ├── services/
│ │ └── twilio-client.ts # Twilio SDK wrapper
│ ├── storage/
│ │ ├── conversation-store.ts # Conversation persistence
│ │ └── message-store.ts # Message persistence
│ ├── tools/
│ │ ├── send-sms.ts
│ │ ├── get-inbound-messages.ts
│ │ ├── create-conversation.ts
│ │ ├── get-conversation-thread.ts
│ │ └── get-message-status.ts
│ └── types/
│ └── models.ts # TypeScript type definitions
└── data/ # SQLite database storage
Development
# Run in development mode
npm run dev
# Build TypeScript
npm run build
# Run tests
npm test
# Run tests in watch mode
npm run test:watch
# Run tests with coverage report
npm run test:coverage
# Type check
npm run lint
Testing
The project includes a comprehensive test suite using Jest with TypeScript support.
Test Structure
tests/
├── setup.ts # Test environment setup
├── config/
│ └── env.test.ts # Environment configuration validation tests
├── services/
│ └── twilio-client.test.ts # Twilio client service tests
├── storage/
│ ├── conversation-store.test.ts # Conversation persistence tests
│ └── message-store.test.ts # Message persistence tests
├── tools/
│ ├── send-sms.test.ts # SMS sending schema validation
│ ├── create-conversation.test.ts # Conversation creation tests
│ ├── get-conversation-thread.test.ts
│ ├── get-inbound-messages.test.ts
│ └── get-message-status.test.ts
└── webhook-server.test.ts # Webhook endpoint tests
What's Tested
- Schema Validation: All MCP tool input schemas are tested for valid/invalid inputs
- Storage Operations: SQLite conversation and message stores (CRUD operations)
- Webhook Handlers: Inbound SMS and status callback endpoints with signature validation
- Environment Config: All environment variable validation rules
- Twilio Client: Client instantiation and method signatures
Running Tests
# Run all tests
npm test
# Watch mode for development
npm run test:watch
# Generate coverage report
npm run test:coverage
Test Environment
Tests use an in-memory SQLite database and mock environment variables configured in tests/setup.ts.
Security
- No Hardcoded Secrets: All credentials are stored in environment variables
- Webhook Validation: All Twilio webhooks are validated using signature verification
- Input Validation: All tool inputs are validated using Zod schemas
- HTTPS Required: Webhook endpoints must use HTTPS in production
Database
The server uses SQLite for storing conversations and messages:
- Conversations: Thread metadata, participants, timestamps
- Messages: SMS/MMS content, delivery status, media URLs
Database file location: ./data/twilio.db (configurable via DATABASE_PATH)
Environment Variables
| Variable | Required | Default | Description |
|---|---|---|---|
TWILIO_ACCOUNT_SID |
Yes | - | Twilio Account SID (starts with AC) |
TWILIO_AUTH_TOKEN |
Yes | - | Twilio Auth Token |
TWILIO_PHONE_NUMBER |
Yes | - | Your Twilio phone number (E.164 format) |
WEBHOOK_PORT |
No | 3000 | Port for webhook server |
WEBHOOK_BASE_URL |
Yes | - | Public HTTPS URL for webhooks |
DATABASE_PATH |
No | ./data/twilio.db | SQLite database file path |
LOG_LEVEL |
No | info | Logging level (debug, info, warn, error) |
AUTO_CREATE_CONVERSATIONS |
No | true | Auto-create conversations for new threads |
ENABLE_AI_CONTEXT |
No | true | Enable AI context generation |
ENABLE_MMS |
No | true | Enable MMS support |
License
MIT
Support
For issues and questions:
- Twilio Documentation: https://www.twilio.com/docs
- MCP Documentation: https://modelcontextprotocol.io
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。