The Autonomous Data

Infrastructure Platform. AI-powered schema intelligence, governance, and evolution.

Forge transforms nested JSON into production-ready dbt models for BigQuery, Snowflake, Databricks, and Redshift—with AI that classifies schemas, recommends governance policies, and evolves with your data.

Stop writing ETL. Deploy infrastructure that understands and manages itself.

{ "event_id": "evt_123", "user": { "id": 42, "traits": { "plan": "pro" } }, "items": [ { "sku": "A-1", "price": 10.00 }, { "sku": "B-2", "price": 5.50 } ] }

TABLE `events_v1` ( event_id STRING, _ingested_at TIMESTAMP ); TABLE `events_v1__user` ( _parent_id STRING, id_hash STRING, traits_plan STRING ); TABLE `events_v1__items` ( _parent_id STRING, sku STRING, price FLOAT64 );

"Data infrastructure that classifies, governs, and evolves itself."

4-Layer AI Intelligence Stack

Forge’s AI understands your data, recommends governance policies, and will soon manage your entire pipeline autonomously.

⚒️ Forge Core

Production

Multi-Warehouse Compiler

The foundation. Transforms deeply nested JSON into production-ready dbt models with automatic normalization and type inference.

• BigQuery, Snowflake, Databricks, Redshift
• 5+ levels deep unnesting
• Automatic dbt docs & lineage

⚔️ Excalibur

Production

Schema Classification

Graph Neural Network that treats your schema as a graph to classify data patterns. Privacy-preserving—field names never leave your environment.

• GraphSAGE GNN + RandomForest
• 89% accuracy across 5 categories
• Privacy-first fingerprinting

🛡️ Pridwen

Production

Governance ML

Hybrid 3-layer system (Rules + ML + Crowd) that detects PII and recommends transformations like hash, mask, and encrypt.

• 15 SQL transformation templates
• Day-1 intelligence out of the box
• Gets smarter with every customer

🐴 Llamrei

Q2 2026

Schema Evolution

Automatically detects legacy API versions and normalizes them to modern golden schemas. Save $200K-$500K per avoided migration.

• 50+ API golden schemas
• Stripe, Salesforce, Shopify support
• Non-destructive transformations

🧙 Merlin

Q4 2026

Autonomous Agent

"Set up my data pipelines and maintain them." LLM-powered agent that understands goals, plans workflows, and self-heals failures.

• Natural language commands
• Multi-step planning & execution
• Learns from outcomes

Data Infrastructure That Manages Itself

❌ Fragile Pipelines: One upstream JSON change breaks your dbt models and crashes your dashboards.

❌ Vendor Lock-in: Custom ETL logic traps you in a single warehouse dialect.

❌ Manual Governance: Engineers waste 30% of their time on PII detection and compliance.

✅ AI-Powered Classification: Excalibur instantly understands your data patterns.

✅ Automatic Governance: Pridwen detects PII and applies transformations—no manual rules.

✅ Universal Compiler: One parse generates optimized SQL for 4 major warehouses.

Infrastructure that understands, governs, and evolves with your data.

🧠 AI Classification

Excalibur’s Graph Neural Network instantly understands your data patterns. Privacy-preserving design means field names never leave your environment.

🛡️ Automatic Governance

Pridwen detects PII and recommends transformations like hash, mask, and encrypt. Day-1 intelligence that gets smarter with every customer.

🔄 Multi-Warehouse

One JSON source compiles to optimized SQL for BigQuery, Snowflake, Databricks, and Redshift. Write once, deploy anywhere.

One JSON Source → Four Optimized Warehouses

Forge transforms nested JSON into query-ready native structs—ready to traverse with dot notation. 60-90% cost savings on repeated queries vs parsing raw JSON every time.

// Raw JSON (parsed every query = expensive)

get_json_object(get_json_object(root, ‘$.patient’), ‘$.name’)

⬇️ Forge Rollup (native struct traversal)

// Query any warehouse with dot notation

frg.root.patient[0].name

Compute Costs

60-90% savings

Query Speed

15x faster

Nesting Depth

5+ levels auto

How is Forge Different?

Traditional ETL tools replicate structured data. Forge understands, governs, and transforms any data—with AI intelligence built in.

Capability	Traditional ETL (Fivetran, Stitch)	Forge
JSON Handling	Loads raw JSON into a single `VARIANT` column. Requires manual parsing.	Automatically unnests nested objects and arrays into clean, queryable tables with proper keys.
Schema Intelligence	No understanding of data semantics. Schemas are just column names.	Excalibur GNN classifies data patterns (payment, customer, inventory) to enable smart defaults.
Data Governance	Manual PII detection. Compliance is your problem.	Pridwen automatically detects PII and recommends hash, mask, or encrypt transformations.
Multi-Warehouse	Separate connectors/config per warehouse. Limited dialect support.	One source → simultaneous output for BigQuery, Snowflake, Databricks, and Redshift with optimized SQL.
Schema Evolution	Pipelines break when fields change. Manual intervention required.	Automatically detects and adapts to schema changes—no downtime. (Llamrei coming Q2 2026)

Focus on Analytics, Not Engineering

Building a robust JSON processing pipeline is a significant engineering effort. Here’s how Forge accelerates your time to insight.

Data Engineering Task	Manual Process (Time Estimate)	With Forge
Initial Schema Discovery	Write scripts to scan data, identify fields, and determine data types. (2-4 hours)	Automatic. Done in minutes during the first run.
Data Classification	Manually review schemas to understand data semantics. (1-2 hours)	Excalibur AI. Instant classification with 89% accuracy.
PII Detection & Governance	Security review, manual tagging, compliance documentation. (1-3 days)	Pridwen AI. Automatic detection and transformation recommendations.
Write Parsing & Unnesting Logic	Develop and debug complex SQL or Python code. (1-3 days)	Automatic. Forge handles all unnesting logic.
Total Time to Value	Days to Weeks	Under 15 Minutes

Security & Compliance: Your Cloud, Your Rules

For organizations with strict data residency and security requirements, Forge Enterprise offers a unique deployment model that keeps you in control.

Requirement	Traditional SaaS	Forge Enterprise
Data Location	Sent to vendor’s cloud	Stays in YOUR cloud
AI/ML Privacy	Your data trains their models	Privacy-first: only structural fingerprints
Security Review Time	3-6 months	2-4 weeks
Works with PII/PHI	Requires extensive compliance	Yes, by architecture

From Black Box to Glass Box: Complete Transparency

Forge isn’t just a parser; it’s a fully-managed, transparent data cataloging system with AI intelligence. Every run generates rich metadata that gives you unprecedented visibility.

📜

Automatic dbt Docs

Forge uses dbt Core to power its transformations. After every run, it generates and hosts a full dbt documentation site for your project.

🔍

Complete Code Transparency

Ever wonder what a transformation tool is actually doing? With Forge, you can inspect the exact SQL code used to generate every table.

🗺️

End-to-End Data Lineage

The generated dbt docs include a complete, interactive DAG showing how data flows from your raw JSON sources to the final tables.

🌐

Write Once, Deploy Anywhere

Forge generates native dbt models for BigQuery, Snowflake, Redshift, and Databricks simultaneously from one JSON source.

🛠️

Full dbt Infrastructure

Get production-ready dbt models executed automatically—no separate dbt Cloud subscription needed (saves $300-500/month).

Get Started with Forge

💰 Transform JSON into AI-governed, query-ready tables for BigQuery, Snowflake, Databricks, or Redshift. Most customers see 3-5x ROI in the first month through compute savings alone.

Start Your Free Trial

🎁

30-Day Free Trial

Full access to test Forge

🧠

AI Features Included

Excalibur + Pridwen in trial

💳

No Credit Card

Start immediately

Pricing Options

Monthly Starter

$299/month

Up to 100M rows, 5 job profiles, all warehouses + AI.

Monthly Professional

$999/month

Up to 1B rows, unlimited profiles, priority support.

Enterprise

Custom pricing

In-VPC deployment, unlimited usage, SLA guarantees. Contact sales.