Product · Integrations

You keep your stack.
You keep control.

Spotonix starts with your semantic and BI assets, SQL, DAX, workspace context, model choice, and warehouse. Onboarding maps the bounded domain into an inspectable path from company meaning to execution.

Book Demo See how it works

What can supply company context

Four ways to bring existing meaning
into the analytical path.

Start with the models, formulas, SQL, names, and relationships your team already maintains. Verify the representative subset during onboarding.

Star-schema or BI-model JSON

Live

Business names, aliases, descriptions, attributes, measures, joins with cardinality, and hierarchies can become context used to interpret questions.

Tested end to end against the reference model and warehouse.

Power BI model structure and DAX

Live

Power BI model structure contributes tables, columns, relationships, measures, and descriptive context. DAX measure formulas are ingested into Intent Algebra so formula meaning can inform interpretation and planning.

Use a representative model during onboarding to verify the DAX constructs and calculated objects the workflow depends on.

dbt-style YAML and semantic assets

Live

YAML and related semantic assets can supply model and column descriptions, semantic columns, SQL, aggregates, joins, metrics, and business names.

Project-specific mappings are handled during onboarding against the assets your team actually maintains.

Existing SQL and comments

Live

SQL written by colleagues can be parsed for analytical structure. Comments inside that SQL can become names and human descriptions Spotonix resolves against.

Dialect and project conventions are verified against a representative corpus during onboarding.

Coverage snapshot: August 2026. The onboarding record names the exact artifacts, formulas, mappings, and warehouse path used for your domain.

Mapped during onboarding

Bring what the domain depends on.
We make the mapping explicit.

Different teams maintain meaning in different systems. Onboarding turns the relevant subset into a dated input, execution, and ownership record—not a blocker list.

Onboarding path

Semantic and BI assets

Bring the dbt, LookML, Cube, AtScale, Snowflake semantic-view, Power BI, or other model artifacts your team maintains. We map the fields and formula classes needed for the bounded domain.

Onboarding path

Warehouse dialect and identity

We verify the SQL dialect, SQLAlchemy driver, customer-VPC network route, credential owner, allowed data slice, and query behavior before the Two-week Pilot begins.

Onboarding path

Model endpoint and credentials

We configure the selected Claude, Gemini, or OpenAI endpoint, credential owner, region, provider terms, and request boundary for the customer environment.

Execution path

Choose the model.
Connect the warehouse.

Model support and database connectivity are configured separately, then verified together against representative questions in the customer VPC.

Claude, Gemini, and OpenAI/Codex

Live

Spotonix supports configured Anthropic, Google Gemini, and OpenAI model paths. The deployment record names the model, endpoint, credential owner, region, provider terms, and data sent with each request.

Choose the model path your team has approved and test the complete surrounding workflow.

Snowflake, Redshift, Databricks, DuckDB, and SQLAlchemy dialects

Live

Spotonix supports Snowflake, Amazon Redshift, Databricks, and DuckDB. Its SQLAlchemy execution layer also allows additional SQLAlchemy-supported warehouse dialects to be configured during onboarding.

We verify dialect, driver, customer-VPC network route, credential mode, allowed data slice, and representative query behavior.

Before the Pilot starts

Turn “Will it connect?” into
five answerable checks.

Data-ready means the bounded source, mappings, warehouse path, model path, and customer-VPC access are explicit. The proof can then start without waiting for a company-wide integration program.

01
Name the input
Identify the exact file, model, SQL corpus, or other artifact—not just the vendor that produced it.
02
Inspect a representative sample
Confirm the fields, formula classes, relationships, comments, and mappings actually present.
03
Map the onboarding inputs
Connect the business names, formulas, relationships, SQL, and physical columns needed for the bounded domain.
04
Confirm the model path
Choose the supported Claude, Gemini, or OpenAI/Codex path, endpoint, model, and credential owner for the deployment.
05
Verify execution in the customer VPC
Test the warehouse dialect, SQLAlchemy driver, network route, credential mode, retained artifacts, and representative queries before Day 1.