Forge, the NoSQL to SQL Compiler
Automates nested JSON flattening for BigQuery and Snowflake better than manual dbt.
Automatically decompose nested JSON in your data warehouse into normalized dbt models. Supports BigQuery, Snowflake, and Databricks
Beats manual SQL macros at flattening nested JSON for dbt pipelines.
Analytics engineers and data teams using dbt with BigQuery or Snowflake
dbt-utils · Fivetran · Airbyte
Forge Core is a deterministic BFS engine that reads a single JSON column (or multi-column table), discovers all nested structures, and generates:
dbt SQL models — one per nested object/array Rollup view — reassembles the full document from normalized tables schema.yml — structural column inventory JSON Schema — standard draft-07 schema of the discovered structure Mermaid ER diagram — table relationship visualization
Automates nested JSON flattening for BigQuery and Snowflake better than manual dbt.
Consistent curl-friendly endpoints for latest assets and a cached +json API are the concrete win here — you can script downloads without wrestling with provider-specific quirks. The modular provider architecture (GitHub, GitLab, Forgejo, cgit) and configurable cache show it was designed for lightweight self-hosting, and the AUR package + examples get you running quickly. It’s useful and pragmatic, but it needs private-repo auth/rate-limit handling and some usage/monitoring features to feel essential.
No AWS credentials needed — parses Terraform plan JSON entirely offline.
Missing primitive: tolerant JSON parser for streaming LLM output, typed and <1KB.
Replaces O(n²) token re-parsing with true O(n) streaming; Vercel SDK does 4K re-parses on 50KB payloads.
Client-side ranking of 50k developers with zero backend—clever constraint engineering.