dbt modelling with Claude
Build and document dbt models without a data hire
Describe the model you need in plain English and Rig drafts the dbt SQL, tests and docs against your real schema, then materialises it into your warehouse.
The problem
Clean, modelled tables are what make every downstream answer trustworthy, but building and maintaining dbt models needs a data engineer. Small teams either go without a model layer or wait weeks for one change.
Built from Rig's building blocks
Rig is not a fixed template. It is a set of building blocks, and this app is one way to assemble them. Take what you need, then shape the workflow around your own pain.
How to build it
- 1
Point Rig at your warehouse
Connect Snowflake or BigQuery read-only. Rig reads your information schema so it models against real tables, not guesses.
- 2
Describe the model
Say what you need, 'a daily revenue fact joining orders, refunds and users', and Rig drafts the dbt SQL with tests and documentation.
- 3
Review the lineage
Check the generated model and its lineage in plain view. Rig validates against your schema so columns and joins are real.
- 4
Materialise and reuse
Push the model into your warehouse so every Rig answer and dashboard runs on it. Iterate in plain English whenever the business changes.
The outcome
A maintained, documented model layer that keeps every downstream number consistent, without hiring a data engineer.
Build your own version
