mcp-siebel-phala

MCP (Model Context Protocol) server for Siebel CRM with HTTP/SSE transport, designed to be deployed on Phala Cloud TEE (Trusted Execution Environment).

Running your Siebel connector inside a TEE ensures that credentials and sensitive data are protected — even the infrastructure provider cannot access them.

Available Tools

Tool	Description
get_account	Get an account by ID
search_accounts	Search accounts by name
get_contact	Get a contact by ID
search_contacts	Search contacts by name or email
get_opportunity	Get an opportunity by ID
create_activity	Create an activity in Siebel
update_record	Update any Siebel record
run_query	Run a query on any Business Object

Requirements

• Docker
• Node.js 20+ (for local development without Docker)

Environment Variables

Copy .env.example to .env and fill in your values:

SIEBEL_URL=https://your-siebel-server:9001
SIEBEL_USERNAME=your_username
SIEBEL_PASSWORD=your_password
SIEBEL_LANG=ENU
PORT=3000

Run locally with Docker

git clone https://github.com/Jeretucu/mcp-siebel-phala.git
cd mcp-siebel-phala
cp .env.example .env
# Edit .env with your Siebel credentials
docker compose up --build

The server will be available at http://localhost:3000.

Endpoints

Method	Route	Description
GET	/sse	SSE connection for the MCP client
POST	/messages	Tool call messages
GET	/health	Health check
GET	/attestation	TDX remote attestation quote (Phala Cloud only)

Deploy to Phala Cloud

Phala Cloud runs your container inside a TEE (Intel TDX), ensuring Siebel credentials are never visible outside the secure environment.

1. Push your image to Docker Hub

docker build -t your-dockerhub-user/mcp-siebel-phala:latest .
docker push your-dockerhub-user/mcp-siebel-phala:latest

2. Create a CVM on Phala Cloud

1. Go to cloud.phala.network and sign in.
2. Click Deploy and paste this docker-compose.yml:

services:
  app:
    image: your-dockerhub-user/mcp-siebel-phala:latest
    container_name: app
    ports:
      - "3000:3000"
    environment:
      - SIEBEL_URL=https://your-siebel-server:9001
      - SIEBEL_USERNAME=your_username
      - SIEBEL_PASSWORD=your_password
      - SIEBEL_LANG=ENU
      - PORT=3000
    volumes:
      - /var/run/tappd.sock:/var/run/tappd.sock
    restart: unless-stopped

3. Click Deploy and wait for the CVM to start.

3. Connect your MCP client

Once running, Phala Cloud assigns a public URL like:

https://xxxxxxxxxxxxxxxx.phala.network

Configure your MCP client (e.g. Claude Desktop):

{
  "mcpServers": {
    "siebel": {
      "url": "https://xxxxxxxxxxxxxxxx.phala.network/sse"
    }
  }
}

4. Verify TEE attestation

curl https://xxxxxxxxxxxxxxxx.phala.network/attestation
# → { "success": true, "quote": { ... } }

The returned quote is a cryptographic proof signed by Intel TDX that your MCP server is running inside a real TEE.

Project Structure

mcp-siebel-phala/
├── src/
│   ├── index.ts              # Express server + SSE
│   ├── siebel-client.ts      # Siebel REST API HTTP client
│   └── tools/
│       ├── accounts.ts
│       ├── contacts.ts
│       ├── opportunities.ts
│       ├── activities.ts
│       ├── records.ts
│       ├── query.ts
│       └── index.ts
├── Dockerfile
├── docker-compose.yml
├── package.json
└── tsconfig.json

Local development (without Docker)

npm install
cp .env.example .env
# Edit .env

# Dev mode (hot reload)
npm run dev

# Build and run
npm run build
npm start