Export and Publishing Surfaces
Semantic models built in Solid can be exported as portable YAML to Snowflake, Databricks, dbt, Tableau, Looker, and more — Solid is a hub, not a closed system.
Semantic models created in Solid can be exported to multiple destinations, making Solid a vendor-neutral semantic hub rather than a closed system. Every export starts from the same source: the Solid YAML representation of your model.
Solid YAML — the portable model
Solid YAML is the canonical, portable format for a semantic model. It captures the full model definition — tables, columns, joins, metrics, custom instructions, and example SQL — in a structured file you can download, version, and move to other systems.
Download Solid YAML from the Versions panel on any semantic model. Every saved version includes a YAML download link, so you always have a snapshot of exactly what was certified at that point. See Semantic Model Versioning for the full version history workflow.
A Solid YAML file includes these asset types:
| Asset type | What it contains |
|---|---|
| Tables & columns | Selected warehouse tables, column metadata, and descriptions |
| Joins / relationships | How tables connect — declared and inferred join paths |
| Metrics | Business metric definitions with their SQL logic |
| Custom instructions | Model-specific context and disambiguation rules |
| Example SQL | Certified query patterns from your query history |
Solid YAML is the foundation for every other export format listed below. When you push to a destination, Solid translates this representation into the target system's native format.
Export destinations
| Export Format | Target System | Use Case |
|---|---|---|
| Solid YAML | Any system | Portable model definition; download from version history or push on save |
| Snowflake Semantic View | Snowflake | Push semantics into Snowflake's native semantic layer for Cortex Analyst and other Snowflake-native consumers |
| DBT YAML | dbt Core / dbt Cloud | Push generated semantic models into a dbt semantic layer repo; dbt Cloud then serves the semantics to BI tools like Sigma |
| Tableau Data Source (.tds) | Tableau | Push validated semantics into Tableau for dashboard consumption |
| LookML YAML | Looker | Push semantics into Looker's semantic layer for Looker-native BI |
| MCP tools | Any MCP-compatible agent | Runtime SQL generation and semantic metadata access for any agent platform |
By warehouse
| Warehouse | What Solid exports |
|---|---|
| Databricks | Solid YAML |
| Snowflake | Solid YAML plus Snowflake Semantic Views — see Enable Snowflake Cortex AI for Solid |
| Other warehouses (Redshift, BigQuery, Fabric, and more) | Solid YAML |
Snowflake is the only warehouse where Solid also creates native semantic objects in the target system. For all other warehouses, Solid YAML is the portable export.
Push on save — your choice
Solid can push model changes to your configured export destinations in two ways:
| Mode | Behavior |
|---|---|
| Automatic push on save | Every time a model is saved, Solid pushes the updated definition to configured destinations |
| Manual export | Solid only pushes when you explicitly trigger an export |
Which mode to use is a customer configuration choice. Teams that want their Snowflake Semantic Views or dbt repo to stay in sync with Solid in real time typically enable push on save. Teams that prefer to review changes before they land in a downstream system can export manually from the model editor or download YAML from version history.
Model ownership
Semantic models are owned by the customer. Solid does not retain exclusive control over exported models — you can push models to your chosen destinations independently, download YAML at any time, and use the definitions in any system that accepts them.
Updated about 1 month ago
