Saison Technology | HULFT

AI, Seamlessly Integrated

Connect your AI applications to enterprise data sources with the HULFT Integrate MCP Adapter. One protocol, unlimited possibilities.

Talk to Sales View Use Cases
20,000+
MCP Servers Available
600+
HULFT Connectors
Bedrock
AWS LLM Platform
Ollama
On-Premise Option

What is the HULFT MCP Adapter?

The HULFT Integrate MCP Adapter is a client component that enables HULFT Integrate to connect to any Model Context Protocol (MCP) server. This allows AI/LLM applications to securely access enterprise data sources through HULFT's existing connector ecosystem.

How It Works

┌─────────────────┐     ┌──────────────────────┐     ┌─────────────────┐
│                 │     │                      │     │                 │
│   LLM/AI App    │────▶│  HULFT MCP Adapter   │────▶│   MCP Servers   │
│                 │     │                      │     │                 │
│  - AWS Bedrock  │     │  - Protocol Handler  │     │  - Salesforce   │
│  - Ollama       │     │  - Auth Management   │     │  - Oracle       │
│  - Claude       │     │  - Tool Orchestration│     │  - PostgreSQL   │
│                 │     │  - HULFT Connectors  │     │  - GitHub       │
│                 │     │                      │     │  - Slack        │
└─────────────────┘     └──────────────────────┘     └─────────────────┘
                                   │
                                   ▼
                        ┌──────────────────────┐
                        │   Enterprise Data    │
                        │                      │
                        │  - Databases         │
                        │  - APIs              │
                        │  - Files             │
                        │  - Legacy Systems    │
                        └──────────────────────┘

Core Capabilities

🔌

MCP Protocol Support

  • ✓ Tools - Execute operations on external systems (CRUD, queries, API calls)
  • ✓ Resources - Read data from files, databases, and APIs
  • ✓ Prompts - Pre-defined prompt templates for common operations
  • ✓ Sampling - Request LLM completions through the server
🔐

Enterprise Security

  • ✓ OAuth 2.1 - Standard enterprise authentication
  • ✓ API Keys - Simple authentication for internal systems
  • ✓ IAM Integration - AWS IAM roles for Bedrock
  • ✓ Credential Vault - Secure credential management
🔄

Data Integration

  • ✓ Multi-Server - Connect to multiple MCP servers simultaneously
  • ✓ Data Transformation - Transform data between systems
  • ✓ Batch & Real-time - Support for both modes
  • ✓ Error Handling - Retry logic and failure recovery
📊

Monitoring & Governance

  • ✓ Audit Logging - Track all MCP operations
  • ✓ Usage Metrics - Monitor API calls and data volume
  • ✓ Rate Limiting - Control API consumption
  • ✓ HULFT Designer - Visual workflow management

Supported LLM Platforms

AWS

AWS Bedrock

Enterprise AI Platform

Supported Models

Claude 3.5 Sonnet, Claude 3 Opus, Amazon Titan, Llama 3, Mistral, Cohere

Integration Features

  • ✓ IAM role-based authentication
  • ✓ VPC endpoints for private connectivity
  • ✓ Bedrock AgentCore MCP server support
  • ✓ CloudWatch logging integration
  • ✓ Guardrails for AI safety

Best For: Enterprise customers with existing AWS infrastructure, compliance requirements, multi-model needs

O

Ollama

Local/Self-Hosted LLMs

Supported Models

Llama 3.2, Mistral, CodeLlama, Phi-3, Gemma 2, Qwen 2.5, DeepSeek

Integration Features

  • ✓ On-premise deployment (air-gapped)
  • ✓ No data leaves your network
  • ✓ REST API compatible
  • ✓ GPU acceleration support
  • ✓ Custom model fine-tuning

Best For: Data sovereignty requirements, air-gapped environments, cost control, development/testing

LLM Compatibility Matrix

Feature AWS Bedrock Ollama
MCP Tool Calling ✓ ✓
Streaming Responses ✓ ✓
Multi-turn Conversations ✓ ✓
Function/Tool Use ✓ ✓
Cloud Hosted ✓ —
On-Premise / Air-Gapped — ✓
No Per-Token Costs — ✓
Enterprise SLA ✓ —

Usage Scenarios

🔄

Scenario 1: AI-Powered ETL Pipeline

Customer wants LLM to analyze data quality issues and automatically fix them during ETL processing.

Flow

  1. HULFT extracts data from source (Oracle, SAP)
  2. MCP Adapter sends sample to LLM for analysis
  3. LLM identifies issues (format, missing values)
  4. Adapter applies corrections via HULFT transforms
  5. Clean data loaded to destination

MCP Servers Used

Oracle MCP PostgreSQL MCP Snowflake MCP

Business Value: Reduce manual data cleansing effort by 60-80%. Catch data quality issues before they impact downstream systems.

💬

Scenario 2: Natural Language Data Queries

Business users query enterprise data using natural language through a chat interface, powered by Bedrock.

Example Queries

  • "Show me all orders over $10K from last month"
  • "Which customers haven't ordered in 90 days?"
  • "Summarize support tickets by category"

MCP Servers Used

Salesforce MCP BigQuery MCP Jira MCP

Business Value: Enable non-technical users to access data without SQL knowledge. Reduce burden on BI team by 50%.

🤖

Scenario 3: AI Agent for IT Operations

Autonomous agent monitors systems and takes corrective actions via MCP servers.

Agent Actions

  • • Monitor database performance via Oracle MCP
  • • Create Jira tickets for anomalies
  • • Post alerts to Slack
  • • Execute remediation scripts via GitHub MCP

MCP Servers Used

Oracle MCP Jira MCP Slack MCP GitHub MCP

Business Value: 24/7 autonomous monitoring. Reduce MTTR (Mean Time to Resolution) by 70%.

🔒

Scenario 4: Air-Gapped AI with Ollama

Financial services firm needs AI capabilities without any data leaving their network.

Architecture

  • • Ollama running on internal GPU servers
  • • HULFT MCP Adapter on-premise
  • • MCP servers for internal databases only
  • • No external network calls

Compliance

SOX PCI-DSS GDPR HIPAA

Business Value: Enable AI capabilities while maintaining strict data sovereignty. Zero data exposure risk.

🔗

Scenario 5: Intelligent CRM-ERP Sync

AI agent keeps Salesforce and Oracle ERP in sync, resolving conflicts intelligently.

Intelligence Layer

  • • LLM detects data conflicts between systems
  • • Applies business rules to resolve conflicts
  • • Learns from human corrections over time
  • • Escalates uncertain cases for review

MCP Servers Used

Salesforce MCP Oracle MCP Slack MCP

Business Value: Eliminate manual data reconciliation. Reduce sync errors by 90%.

Compatible MCP Servers

The HULFT MCP Adapter can connect to any MCP-compliant server. Here are the most common enterprise servers:

🗄️ Databases

💼 Business Apps

🛠️ Developer Tools

💬 Communication

Note: With 20,000+ MCP servers available, customers can connect to virtually any system. Custom MCP servers can also be built for proprietary systems using the official SDKs.