MotherDuck: Brainforge warehouse overview
Last updated: 2026-04-09
Audience: Data / analytics engineers working in Brainforge’s MotherDuck org (Default, Omni, marts).
Purpose
High-level map of databases and schemas you are likely to see, how they relate to pipelines and playbooks, and where client-specific detail lives. This is not a full data catalog; use SHOW ALL TABLES and duckdb_tables() in MotherDuck for ground truth.
Where documentation lives
| Location | Contents |
|---|---|
| MotherDuck CLI setup | Install DuckDB CLI, MOTHERDUCK_TOKEN, verify connection |
| Default export load playbook | Step-by-step: Default product CSV → my_db → raw_export |
| Default export pipeline | Export bundle shape, raw_* tables, SQL script names (brainforge-bi repo) |
| This file | Org-level layout and pointers |
Databases (typical)
Names may evolve; always run SHOW DATABASES; after duckdb "md:".
| Database | Role |
|---|---|
my_db | Staging and scratch for Default product CSV loads; date-window schemas like stg_YYYYMMDD_YYYYMMDD. |
raw_export | main.raw_* tables from Default product exports (teams, events, meetings, submissions, etc.) consumed by dbt/Omni. See Default export pipeline. |
raw_data | Additional ingested sources (not the Calendly-style product CSV pipeline). Example: raw_data.plain — Plain (support) objects such as threads, messages, companies (sync/connector managed outside this repo’s export scripts). |
prod, dev, cicd | Environment or project databases as used by delivery; confirm ownership with the data team before changing. |
Product export pipeline vs other data
- Default product operational exports (large CSV folders from Default) → documented load path:
my_db→raw_export. SQL in the brainforge-bi repo underscripts/(see pipeline doc). - Plain and other SaaS sources may land in
raw_data(or elsewhere) via separate ETL/sync tools; they are not described by theraw_exportCSV playbook.