Salesforce MCP Server for Heroku

A Model Context Protocol (MCP) server that enables AI agents to interact with Salesforce CRM data.

Features

• SOQL Queries - Execute any SOQL query against your Salesforce org
• Object Metadata - Discover fields and relationships for any object
• Record Management - Create, update, and delete records
• Heroku Native - Designed for Heroku deployment with MCP process support

Available Tools

Deployment

One-Click Deploy

Click the "Deploy to Heroku" button above, then configure your Salesforce credentials.

Manual Deploy

git clone https://github.com/dsouza-anush/mcp-salesforce
cd mcp-salesforce
heroku create my-salesforce-mcp
git push heroku main

Configure Credentials

heroku config:set SF_USERNAME=your-username@example.com
heroku config:set SF_PASSWORD=your-password
heroku config:set SF_SECURITY_TOKEN=your-security-token
heroku config:set SF_LOGIN_URL=https://login.salesforce.com

For sandbox orgs, use https://test.salesforce.com as the login URL.

Register with Heroku AI

After deployment, attach your Heroku AI model:

heroku addons:attach your-main-app::INFERENCE -a my-salesforce-mcp

The mcp-salesforce process is now available via /v1/agents/heroku.

Usage Example

from openai import OpenAI

client = OpenAI(
    base_url=os.getenv("INFERENCE_URL") + "/v1",
    api_key=os.getenv("INFERENCE_KEY")
)

response = client.chat.completions.create(
    model=os.getenv("INFERENCE_MODEL_ID"),
    messages=[
        {"role": "user", "content": "Find all opportunities closing this quarter"}
    ],
    extra_body={
        "heroku": {
            "mcp_servers": ["mcp-salesforce"]
        }
    }
)

Environment Variables

License

MIT