santoshprolocity/salesforce-mcp-server

Salesforce MCP Server

A middleware bridge allowing conversational AI assistants like Claude to interact with Salesforce. This server acts as an intermediary that translates natural language requests into Salesforce API operations, enabling a seamless conversational interface to your Salesforce data and metadata.

? Bridge between AI and Salesforce

The Salesforce MCP (Metadata API/Conversational Programming) Server enables you to:

• Ask your AI assistant to connect to Salesforce and perform actions
• Use natural language to query, create, update and analyze Salesforce data
• Navigate between records with conversational commands
• Generate metadata components through natural language descriptions
• Execute Apex code and automation through simple prompts

? Example Use Cases

You: "Connect to my production Salesforce org"
[System performs OAuth authentication]

You: "Query the top 10 accounts by created date"
[System returns account list]

You: "Open the first account in that list and show me its related contacts"
[System retrieves and displays the related contacts]

You: "Generate a trigger that updates opportunities when accounts change"
[System generates and displays Apex code]

You: "Create a new field on Contact called Training Status"
[System creates new metadata]

? Core Features

Authentication & Connectivity

• Multi-Org Authentication - Connect to production, sandbox, and scratch orgs
• OAuth 2.0 Flow Support - Web server, JWT, and device authentication flows
• Session Management - Secure storage and refresh of access tokens
• Connection Profiles - Save and switch between different org connections

Data Operations

• Conversational Querying - Translate natural language to SOQL
• Object Navigation - Browse related records with simple commands
• Record Retrieval - Get records by ID, name or other unique identifiers
• Record Creation/Updates - Create and update records via conversation
• Record Search - Find records across multiple objects with SOSL

Metadata Operations

• Natural Language Metadata Creation - Generate fields, objects, and components
• Metadata Retrieval - View and analyze org metadata with simple requests
• Apex Code Generation - Create Apex classes, triggers, and more via conversation
• Metadata Deployment - Deploy components to orgs with natural commands

AI Integration

• Context Awareness - Maintain context between requests
• Command Recognition - Parse and understand Salesforce-specific commands
• Intent Detection - Map natural language to appropriate Salesforce API calls
• Entity Extraction - Identify objects, fields, and values in requests
• Response Formatting - Present Salesforce data in a conversational manner

?? Installation

git clone https://github.com/santoshprolocity/salesforce-mcp-server.git
cd salesforce-mcp-server
pip install -r requirements.txt
python app.py

? Connecting to AI Assistants

The server exposes RESTful endpoints that can be called by AI assistants:

1. Authentication: /api/auth/* - Endpoints for connecting to Salesforce orgs
2. Data Operations: /api/data/* - Endpoints for data querying and manipulation
3. Metadata Operations: /api/metadata/* - Endpoints for metadata management
4. Natural Language Processing: /api/nlp/* - Endpoints for parsing requests
5. AI Context Management: /api/context/* - Endpoints for managing conversation context

? Getting Started

1. Configure your environment variables in .env
2. Start the server with python app.py
3. Connect your AI assistant to the server APIs
4. Start asking your AI to interact with Salesforce!

? Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

? License

This project is licensed under the MIT License - see the LICENSE file for details.