Getting Started with Forge

From GCP Marketplace to Production dbt in 5 Minutes

Prerequisites

GCP Project with billing enabled
BigQuery dataset with JSON data
Service Account with BigQuery Data Editor role
5 minutes of your time

1 Subscribe on GCP Marketplace

Visit GCP Marketplace
Search for "Forge Data Platform"
Choose your plan: Starter ($299/mo), Professional ($999/mo), or On-Demand ($1,999)
Click Subscribe and confirm billing

2 Sign In to Forge Portal

After subscribing, click "Manage on Provider" or go directly to the Forge Portal. Sign in with Google SSO using the same account from GCP Marketplace.

3 Upload Service Account

In Forge Portal, go to Organization Settings
Click "Add Service Account"
Upload your service account JSON key file

4 Create Job Profile

Click "Create Job Profile" and provide:

Source Project ID
Source Dataset & Table
Target Dataset (where models will be created)

5 Run Your First Job

Click "Run Job" next to your profile. Watch real-time progress. Job completes in 2-5 minutes depending on data size.

Tip: Enable "Dry Run" to preview schema without processing all data.

Understanding Dry Run Mode

Before processing your entire dataset, use Dry Run to preview the transformation without materializing tables:

Samples 10,000 rows (configurable) to discover schema
Previews the tables and views that will be created
Estimates "amplification factor" (row explosion from unnesting arrays)
Verifies your job configuration without processing millions of rows

What You Get in Dry Run:

Included

JSON schema preview
Table structure diagram
Estimated row counts per table
Nesting depth analysis

Not Included

Materialized tables in BigQuery
Downloadable dbt artifacts
Full data processing
Production-ready models

Best Practice: Always run Dry Run first on new datasets, especially if you're unsure about nesting depth or have large tables (100M+ rows).