# SemanticOps MCP

> SemanticOps MCP, formerly MCP Engine, is an AI-native bridge between language models and Power BI semantic models. Control your data models with natural language through Claude, Copilot, and any MCP-compatible client.

## What is SemanticOps MCP?

SemanticOps MCP is a local tool that connects AI assistants to Power BI using the Model Context Protocol (MCP). It was formerly called MCP Engine. It lets you query, create, and modify semantic models using natural language instead of clicking through menus or writing scripts.

SemanticOps Test Runner is a desktop app included with SemanticOps Pro. It provides guided semantic-model testing backed by SemanticOps MCP. On Windows it registers as a Power BI Desktop External Tool. On macOS it connects to Power BI Service.

## Who It's For

- **Power BI Developers**: Accelerate semantic model development with AI-assisted DAX authoring, bulk operations, and natural language queries
- **Data Analysts**: Query and explore models using plain English, no DAX expertise required
- **BI Consultants**: Manage multiple client models efficiently, standardize patterns, and document models automatically
- **Enterprises**: Govern Power BI at scale with audit trails, security role management, and centralized model control

## Key Capabilities

- Query 15 object types (tables, measures, columns, relationships, etc.)
- Create and modify measures, calculated columns, KPIs, calculation groups
- Run DAX queries and analyze performance with query plans
- Manage security roles and row-level security
- Bulk operations with dry-run preview
- Transaction history with rollback support
- PII masking for data protection (Pro/Enterprise)
- SemanticOps Test Runner desktop app for guided semantic model unit testing (Pro)
- Test packs, regression snapshots, DAX assertions, RLS/OLS validation, and JUnit/Markdown exports (Pro)

## How It Works

1. AI client (Claude, Copilot, etc.) sends natural language request
2. SemanticOps MCP parses intent and translates to TOM operations
3. Operations execute locally via ADOMD.NET
4. Structured results return to AI client

All processing happens locally on your machine. No cloud, no data collection.

## Supported Environments

- Power BI Desktop (Windows)
- Power BI Service via XMLA endpoint (Premium/Fabric capacity)

## Pages

- [Home](https://www.semanticops.dev/): Product overview and quick start
- [Getting Started](https://www.semanticops.dev/getting-started): Installation guides for Claude Desktop, VS Code extension, and manual setup
- [Features](https://www.semanticops.dev/features): Complete list of 15 tools and capabilities
- [Test Runner](https://www.semanticops.dev/test-runner): SemanticOps Test Runner desktop app for Power BI semantic model testing
- [Compare](https://www.semanticops.dev/compare): SemanticOps MCP vs DAX Studio, Tabular Editor, Measure Killer
- [Pricing](https://www.semanticops.dev/pricing): Free, Pro ($349/year), and Teams plans
- [FAQ](https://www.semanticops.dev/faq): How it works, capabilities, privacy, and comparisons

## Quick Start

```
claude mcp add semanticops -- npx -y mcp-engine@latest
```

## Links

- [GitHub](https://github.com/maxanatsko/mcp-engine-public)
- [Contact](mailto:support@semanticops.dev)