Salesforce Basic MCP Server

A simple Model Context Protocol (MCP) server that provides AI assistants with tools to query Salesforce data.

Features

• ✅ Get Account by ID

• ✅ Search Accounts by name

• ✅ Get recent Opportunities

Setup

1. Install Dependencies

2. Configure Salesforce Credentials

Create a .env file in the project root:

How to get your Security Token:

1. Log into Salesforce

2. Click your profile icon → Settings

3. On the left sidebar: Personal → Reset My Security Token

4. Check your email for the token

3. Run the Server

4. Test with MCP Inspector

Available Tools

get_account

Get a Salesforce Account by ID.

Parameters:

• account_id (string): The Salesforce Account ID

Example:

search_accounts

Search for Accounts by name.

Parameters:

• name (string): The account name to search for

Example:

get_recent_opportunities

Get the most recent Opportunities.

Parameters:

• limit (integer, optional): Number to return (default: 5, max: 20)

Example:

Connect to Claude Desktop

Add to your Claude Desktop config file:

Mac: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config. json

Security Notes

⚠️ Never commit

• The .gitignore file is configured to exclude it

• Use environment variables in production

Troubleshooting

"Authentication Failed"

• Double-check username, password, and security token

• Verify your IP is not blocked in Salesforce

• For sandboxes, use SF_DOMAIN=test

"Module not found"

• Make sure virtual environment is activated

• Run uv add "mcp[cli]" simple-salesforce python-dotenv

Next Steps

• Add more Salesforce objects (Contacts, Cases, etc.)

• Implement create/update operations

• Add error logging

• Add data validation