# Open semantic layer

> Your metrics shouldn't be locked inside a single tool. Query your Holistics semantic layer from any application.

## Your metrics, everywhere

Traditional BI tools lock your metric definitions inside their platform. Define "revenue" in one tool, and you can't use that same logic in a Python script or feed it to an AI agent.

Holistics takes a different approach: **your semantic layer is open**. Once you define a metric in Holistics, you can query it from anywhere - notebooks, internal applications, data pipelines, or any system that makes HTTP requests. One definition, unlimited consumption.

## What makes it open

An "open" semantic layer isn't just about having an API. It's about giving you full ownership and control over your business logic through three pillars:

### Code-based definitions

Your metrics are defined in [AML (Analytics Modeling Language)](/reference/aml/) - human-readable code that lives in your repository, not hidden inside a proprietary database.

This means your semantic layer is:

- **Portable** - Move between environments or tools without losing your work
- **Reviewable** - Use code review workflows for metric changes
- **Transparent** - Anyone can read and understand how metrics are calculated

### Version controlled

Because your semantic layer is code, you get the full power of [Git version control](/docs/git-version-control):

- **History** - See who changed what and when
- **Branching** - Test metric changes in isolation before merging
- **Rollback** - Revert problematic changes instantly
- **Collaboration** - Multiple team members can work on different parts simultaneously

### Programmatically accessible

Query your metrics from anywhere via [API](/api/v2/query-data) and [CLI](/docs/cli):

- **API** - HTTP endpoints let any application fetch governed metrics
- **CLI** - Local development tools integrate with your existing workflows
- **CI/CD** - Validate metric definitions automatically before deployment

Together, these ensure you're never locked in. Your business logic stays yours.

## Why this matters

When your semantic layer is open, you get:

- **Single source of truth** - One metric definition serves dashboards, notebooks, internal apps, and AI agents
- **No vendor lock-in** - Your business logic is accessible via API, not trapped in a proprietary format
- **Governed flexibility** - Centralized definitions with decentralized access means consistency without bottlenecks

This means you can invest in building a rich semantic layer in Holistics without worrying about future flexibility. Your work stays accessible regardless of how your data stack evolves.

## What you can build

With programmatic access to your Holistics datasets, you can extend your metrics beyond dashboards:

| Use case | Description |
|----------|-------------|
| **Enrich analysis in notebooks** | Pull metrics into Jupyter, Python, or R for ad-hoc analysis that goes beyond what dashboards offer |
| **Power internal applications** | Serve trusted numbers to operational tools, admin panels, or customer portals |
| **Connect other BI tools** | Query from Metabase, Looker Studio, or any tool that can make HTTP requests |
| **Unit test metrics in CI/CD** | Validate metric definitions programmatically before deploying changes |
| **Enable AI agents** | Feed governed metrics to LLMs and AI assistants using the [MCP Server](/docs/ai/mcp-server) |

**[Get started with the API tutorial →](/api/v2/query-data)**

## Related capabilities

- **[MCP Server](/docs/ai/mcp-server)** - Let AI agents query your semantic layer directly
- **[Embedded Analytics](/embedded/)** - Embed full dashboards in your product
- **[Validation API](/docs/continuous-integration/validation-api)** - Validate AML changes in CI/CD pipelines