sergiocbx/mcp-salesforce repository overview

⁠Salesforce MCP Server

A production-ready Model Context Protocol (MCP) server that enables AI assistants to interact with Salesforce data and automation through dynamically configurable tools.

⁠What is Salesforce MCP Server?

Salesforce MCP Server acts as a bridge between AI assistants (Claude, ChatGPT, Gemini, etc.) and your Salesforce org. It implements the MCP 2025-03-26 specification using Streamable HTTP transport.

Key Features:

• Execute Apex REST endpoints via AI commands
• Configure custom tools without additional code
• OAuth 2.0 + PKCE authentication
• Multi-org support
• Session management with automatic cleanup

⁠How to use this image

⁠Quick Start

docker run -d \
  -p 3000:3000 \
  --name salesforce-mcp \
  -e MCP_PUBLIC_URL="https://your-server.com" \
  -e MCP_SESSION_SECRET="$(openssl rand -hex 32)" \
  -e SF_CLIENT_ID="your_consumer_key" \
  -e SF_CALLBACK_URL="https://your-server.com/auth/callback" \
  -e LICENSE_KEY="your_license_key" \
  sergiocbx/mcp-salesforce:latest

⁠Using Docker Compose

services:
  mcp-salesforce:
    image: sergiocbx/mcp-salesforce:latest
    ports:
      - "3000:3000"
    environment:
      - MCP_PUBLIC_URL=https://your-server.com
      - MCP_SESSION_SECRET=your_32_character_secret_minimum
      - SF_CLIENT_ID=your_consumer_key
      - SF_CALLBACK_URL=https://your-server.com/auth/callback
      - LICENSE_KEY=your_license_key
    restart: unless-stopped

⁠Environment Variables

⁠Required

⁠Optional

⁠Salesforce Setup

Before using this image, you need to:

1. Install the managed package in your Salesforce org:

   /packaging/installPackage.apexp?p0=04tU6000000VN6nIAG
   

   Copy

2. Create a Connected App with:

   • OAuth enabled
   • PKCE required
   • Scopes: full, refresh_token, offline_access

⁠Endpoints

⁠Documentation

Full documentation available at: GitHub Repository⁠

⁠Tags

• latest - Latest stable release
• x.y.z - Specific version

⁠Architecture

Supports linux/amd64 and linux/arm64 platforms.