Provides natural language querying capabilities for Salesforce data with intelligent SOQL/SOSL generation, cross-object search, data manipulation operations (DML), and automated field resolution. Optimized for wholesale distribution systems with support for inventory management, order processing, and supply chain operations.
Orion Salesforce MCP Server
A powerful and intelligent Salesforce data interaction server designed specifically for wholesale distribution systems. This MCP (Model Control Plane) server provides natural language querying capabilities with advanced SOQL/SOSL generation, intelligent field resolution, and comprehensive data manipulation features.
🌟 Key Features
Intelligent Data Querying
Natural Language to SOQL: Convert plain English questions into optimized Salesforce queries
Smart Object & Field Resolution: Automatically map business terms to Salesforce API names using configurable synonyms
Cross-Object Search (SOSL): Search across multiple objects simultaneously with intelligent result aggregation
Relationship Traversal: Automatically handle lookups and master-detail relationships
Time-Sensitive Filtering: Intelligent date range detection and application
Data Manipulation (DML)
Controlled DML Operations: Configurable Insert, Update, Upsert, Delete, Undelete, and Merge capabilities
Feature Flags: Granular control over DML operations with org-specific configuration
Safety Guardrails: Built-in limits, confirmation requirements, and bulk operation controls
Permission Respect: Honor Salesforce field-level security and object permissions
Hybrid AI Responses
Data-Driven Answers: Combine actual Salesforce data with LLM knowledge for contextual responses
Intelligent Fallbacks: Graceful handling when data is unavailable or insufficient
Business Context: Understand wholesale distribution terminology and processes
Enterprise-Ready Architecture
Multi-Org Support: Configure different Salesforce orgs with unique settings and permissions
Robust Error Handling: Comprehensive Salesforce API error handling with clear user feedback
Security First: Token management, PII redaction, and configurable access controls
Performance Optimized: Intelligent caching, query optimization, and efficient API usage
🚀 Quick Start
Prerequisites
Node.js 18+
Salesforce org with API access
Valid Salesforce access tokens
Installation
Clone and Install
git clone https://github.com/jasonjuela/orion-salesforce-mcp-server.git cd orion-salesforce-mcp-server npm installConfigure Salesforce Tokens
cp data/secrets/tokens.example.json data/secrets/tokens.jsonEdit
data/secrets/tokens.json
with your Salesforce credentials:{ "default": { "access_token": "your_access_token_here", "instance_url": "https://your-instance.salesforce.com", "token_type": "Bearer" } }Customize Organization Profile
Edit
data/configs/default.json
to match your Salesforce org structure:{ "orgId": "your-org-name", "namespaces": ["your_namespace__"], "objectSynonyms": { "owsc__Item__c": ["wine", "product", "inventory"], "owsc__Order__c": ["order", "purchase order", "po"] } }Start the Server
npm startServer runs on
http://localhost:3018
📖 Usage Examples
Basic Querying
Cross-Object Search
Test Interface
Open http://localhost:3018
in your browser for a simple web interface to test queries.
⚙️ Configuration
Organization Profiles (data/configs/default.json
)
DML Operations
To enable data manipulation:
Set
features.dmlOperations.enabled: true
Enable specific operations (
insert
,update
, etc.)Configure safety guardrails
Specify allowed/blocked objects in
dmlRestrictions
🏗️ Architecture
Core Components
Planner Service (
src/services/planner.js
): Intent detection and query generationSalesforce Service (
src/services/salesforce.js
): API client with retry logicIntelligent Resolver (
src/services/intelligentResolver.js
): Object and field mappingStream Generator (
src/routes/generateStream.js
): Main orchestration and LLM integration
Data Flow
Intent Detection: Analyze user question for search type and entities
Object Resolution: Map business terms to Salesforce objects using synonyms
Field Discovery: Use Describe API to find relevant fields
Query Generation: Build optimized SOQL/SOSL queries
Data Retrieval: Execute queries with error handling and retries
Response Generation: Combine data with LLM for intelligent answers
🔧 API Endpoints
Primary Endpoints
POST /generate/stream
- Main query interface with streaming responsesGET /auth/login
- Salesforce OAuth initiation (placeholder)GET /auth/callback
- OAuth callback handler (placeholder)GET /search/capabilities
- Check available search capabilitiesPOST /search/cross-object
- Direct SOSL search endpoint
Response Format
Streaming responses include:
Query plan (SOQL/SOSL)
Data retrieval status
Record count
Formatted results
LLM-generated insights
🛡️ Security Features
Token Management: Secure credential storage and rotation
PII Redaction: Configurable personally identifiable information filtering
Object Whitelisting: Control access to specific Salesforce objects
Field-Level Security: Respect Salesforce permissions
Query Limits: Prevent expensive or dangerous operations
Feature Flags: Granular control over dangerous operations like DML
🎯 Wholesale Distribution Focus
Optimized for wholesale/distribution use cases:
Inventory Management: Track items, lots, containers, and locations
Order Processing: Handle purchase orders, sales orders, and fulfillment
Supply Chain: Manage vendors, warehouses, and shipping
Financial Operations: Support invoicing, payments, and accounting
Business Intelligence: Aggregate data for reporting and analytics
🚦 Production Considerations
Performance
Implement response caching for frequent queries
Use connection pooling for Salesforce API calls
Add request rate limiting
Monitor API usage limits
Security
Implement proper OAuth flows
Add request authentication/authorization
Enable HTTPS in production
Regular token rotation
Audit logging
Monitoring
Add comprehensive logging
Implement health checks
Monitor Salesforce API limits
Track query performance metrics
🤝 Contributing
Fork the repository
Create a feature branch (
git checkout -b feature/amazing-feature
)Commit changes (
git commit -m 'Add amazing feature'
)Push to branch (
git push origin feature/amazing-feature
)Open a Pull Request
📄 License
This project is licensed under the MIT License - see the LICENSE file for details.
🆘 Support
For questions, issues, or feature requests:
Open an issue on GitHub
Check existing documentation
Review configuration examples
Built for intelligent Salesforce data interaction with wholesale distribution in mind.
This server cannot be installed
hybrid server
The server is able to function both locally and remotely, depending on the configuration or use case.
Enables natural language querying and data manipulation of Salesforce orgs, with intelligent SOQL/SOSL generation, cross-object search, and enterprise-ready features optimized for wholesale distribution systems.
Related MCP Servers
- -securityFlicense-qualityEnables users to upload retail data, analyze trends, optimize inventory, and forecast sales using AI-powered insights, acting as a senior supply chain expert.
- -securityFlicense-qualityEnables LLMs to interact with DataForSEO and other SEO APIs through natural language, allowing for keyword research, SERP analysis, backlink analysis, and local SEO tasks.Last updated -41053
- AsecurityAlicenseAqualityA server implementation that enables LLMs to interact with Salesforce data through SOQL queries, SOSL searches, and API operations including retrieving, creating, updating, and deleting records.Last updated -10MIT License
- AsecurityAlicenseAqualityProvides flexible access to Oracle databases for AI assistants like Claude, supporting SQL queries across multiple schemas with comprehensive database introspection capabilities.Last updated -65MIT License