Hello team!
Over the past few months, I’ve been getting pretty tired of Claude Code constantly getting lost when working with dbt 
— Agent confuses schemas and tables (prod vs dev, alias vs filename)
— Hallucinates non-existent columns and wrong data types
— Doesn’t know dependencies between models
— Spends forever scanning .sql files and BigQuery for metadata
I have ~1000 models, custom naming patterns, aliases, and defer workflow. So, I’ve built a solution.
dbt-meta — CLI for instant metadata extraction from dbt artifacts.
meta schema dim_customers → project.analytics.customers
meta columns -j orders → [{“name”: “order_id”, “type”: “INT64”}, …]
meta parents -a model → all upstream dependencies
and many other commands…
Most commands ~250ms, not 5-10 tool calls at 2-3 seconds each
pip install dbt-meta
GitHub - Filianin/dbt-meta: AI-first CLI for dbt metadata extraction
Give it a try — feedback welcome!
