Semantic Layer

Stop hand-coding semantic layers

Semantic layers are the interface between AI tools and your data. Other tools make you define every relationship manually. Squish auto-generates semantic models from discovered relationships and exports to dbt, Snowflake, Databricks, and more.

Start free
Talk to us
metricflow_model.yml
Export To
dbt MetricFlow
Snowflake Cortex
Databricks Genie

Why this matters

AI needs semantic context to be accurate

Snowflake Intelligence and Databricks Genie can answer questions about your data. But only if they understand how tables relate. Squish builds that context automatically so AI-generated queries are actually correct.

The only tool that generates semantic layers

Auto-Generated, Not Hand-Coded

Other semantic layer tools require you to manually define every relationship, dimension, and measure. Squish generates definitions from discovered relationships. You review and refine, not start from scratch.

AI-first architecture

Built for AI Agents

The future of BI is AI agents querying semantic layers. Snowflake Intelligence, Databricks Genie, and Looker Conversational Analytics all need semantic context to be accurate. Squish builds that context automatically.

One model, multiple destinations

Cross-Platform Export

Define once, export everywhere. Your semantic model works in dbt, Snowflake, Databricks, and more. As the Open Semantic Interchange standard matures, we support it too.

Export anywhere

One semantic model, multiple destinations

dbt MetricFlow

Available

Export directly to dbt Semantic Layer format. Compatible with dbt Cloud and dbt Core.

Snowflake Cortex

Available

Generate semantic views that Cortex Analyst can consume directly for natural language queries.

Databricks Unity

Available

Create Metric Views in Unity Catalog. Give Databricks Genie the context it needs.

LookML

Available

Generate LookML files with relationships, dimensions, and measures defined.

Cube.dev

Available

Export to Cube semantic layer format for universal BI connectivity.

Power BI TMDL

Available

Generate Tabular Model Definition Language files for Power BI datasets.

AWS QuickSight

Available

Export semantic definitions for QuickSight Q natural language queries.

ThoughtSpot TML

Available

Generate ThoughtSpot Modeling Language for search-driven analytics.

Open Semantic Interchange

Available

Export using the open OSI standard for cross-platform semantic model portability.

How it works

From discovery to export

01

Discover relationships

Squish scans your databases and finds relationships using statistical analysis, not just foreign keys.

02

Generate semantic model

We auto-generate entities, relationships, dimensions, and measures from your discovered schema.

03

Review and refine

Add business context, adjust names, and validate the model. The structure is done, you add meaning.

04

Export to your tools

One click to export to dbt, Snowflake Cortex, Databricks, or any supported format.

Squish vs. manual semantic layers

Other tools require you to manually define everything. We generate, you refine.

CapabilitySquishdbtCubeAtScale
Auto-discover relationships
Cross-database support
Confidence scoring
Continuous monitoring
Export to multiple formats

Let AI understand your data

Auto-generate semantic models from discovered relationships. Export to dbt, Snowflake, Databricks, and more. Give your AI tools the context they need to be accurate.

Start building