Why we’re building an open semantic layer at Lightdash

The business intelligence world has a problem. Every BI tool speaks its own language. Every platform defines metrics differently. And every time you want to move or integrate, you start from scratch. This has been by design. BI tools traditionally wanted to keep your business logic in their platform in a way that makes it non-trivial to migrate or use elsewhere.

‍

We think there's a better way.

‍

That's why we've open-sourced our Lightdash semantic layer. It’s a simple way to define business metrics that lives alongside your dbt project, fully version-controlled and configured as code. It's the same layer that powers our own platform, and now it's available for anyone to use across any tool.

‍

The problem we're all facing

‍

A quick recap on why you need a semantic layer to start with - you're in a meeting and someone asks about monthly recurring revenue. Three people share three different numbers. Each one is "right" according to their tool, but nobody trusts any of them.

‍

Sound familiar? This "metrics drift" wastes countless hours and erodes confidence in data. Teams argue about definitions instead of acting on insights. Data teams drown in repetitive requests. Business users either wait days for answers or create their own (often wrong) reports.

‍

All of this because business logic gets locked inside proprietary tools. When your metrics live in tools like Tableau, Power BI, or Looker, you're stuck. Want to try a new visualization tool? Start over. Need to integrate with a data science platform? Start again from scratch. Want to leverage it to build AI agents? Nope.

‍

An open standard for semantic layers

‍

A semantic layer translates raw data into business concepts like revenue, active users, or churn rate. It's the literal logic layer of your business.

‍

However most semantic layers create silos instead of solving them. They're fragmented, closed, or built to serve a single product. That's backwards. The entire benefit of a semantic layer should be that it’s widely accessible and ideally, built on an open standard.

‍

An open semantic layer changes this paradigm. It decouples semantics from proprietary tools, and that changes everything. Here's what becomes possible:

‍

• True consistency everywhere. Define "customer lifetime value" once, and it means exactly the same thing in Excel, Tableau, or when asking an AI assistant.
• Freedom to choose your tools. Your semantic definitions become portable. You can experiment with new platforms without rebuilding your entire knowledge base.
• AI that gets your business. Generative AI tools rely on semantic layers to interpret natural language questions and convert them into accurate SQL queries. Open standards give AI systems the business context they need for trustworthy answers.
• Real-time insights without data movement. Everything works with your existing warehouse, so you get speed without copying data everywhere.

‍

Built for builders: The Lightdash semantic layer

‍

We dogfood our semantic layer every day. It’s used by thousands of teams daily as it underpins the entire Lightdash platform, and it was built for analytics. Our semantic layer also powers our Embedding service, our REST APIs and even our Lightdash MCP for building AI Agents.

‍

Our semantic layer lives as YAML files right alongside your dbt project. That means:

• Version control comes as standard (no more "I think we changed this last week").
• Works with Cursor/Co-pilot for fast development and fully supported with CI/CD platforms.
• It’s human-readable while still being AI-ready.
• No extra proprietary syntax to learn. If you know SQL, you’re already an expert.

‍

We’ve created a number of ways that you can interact with the Lightdash semantic layer:

• The Lightdash UI for all the BI work you need.
• REST API for powering your own applications or embedded services.
• Python Client for leveraging the semantic layer directly inside data science tools like notebooks.
• Lightdash MCP Server for providing tools to your own custom AI agents.
• Google Sheets Extension - yes we know, spreadsheets!

‍

We’re also excited to  build out more first-party integrations over time, and we’re already seeing customers build their own custom connectors and AI Agents across their teams.

‍

The bigger picture

‍

AI and BI are blending fast. And the tools that win will be the ones that make it easy for people and AI to work with data together. That requires a shared language for business metrics; one that's open, community-driven, and puts you in control of your own definitions.

‍

Open standards reduce dependence on any single vendor and provide flexibility to adopt new tools and technologies as they emerge. As the data landscape keeps evolving (and it will), you want to be ready for whatever comes next.

‍

Help us define the standard

‍

The future of BI should be collaborative, not competitive. It should be about making data more accessible, not more locked down. And it should put you in control of your own business definitions.

‍

Join us in building an open semantic layer. Next, I’ll share some hands-on tutorials on how you can get started with the Lightdash semantic layer and showcase some awesome things our community are already using it for.

‍

Ready to try our open semantic layer? Check out our documentation or talk to our team to see how it can work with your existing dbt project.

‍