April 29, 2026
Rocky road to the comments
Show HN: Rocky – Rust SQL engine with branches, replay, column lineage
This new data tool wowed some people — and instantly made skeptics smell hype
TLDR: Rocky is a new open-source tool that promises safer, easier control over company data workflows without replacing existing cloud services. Commenters were split between praising its clever tracking features and accusing the launch pitch of hype, fuzzy claims, and a little anti-rival swagger
A new project called Rocky just strutted onto Hacker News promising to be the “trust system for your data” — basically, a tool meant to help teams keep their business data pipelines from quietly breaking. It boasts flashy features like safe experimentation, replaying past runs, tracking where every spreadsheet-style column came from, and even catching mistakes before anything goes live. In plain English: Rocky wants to be the traffic cop for messy company data, while still letting teams keep big cloud services like Snowflake or Databricks.
But the real fireworks were in the comments. One camp was genuinely impressed, especially by Rocky’s ability to trace a single piece of data before anything runs. One commenter said most data-tracing tools feel like “archaeology after the fact,” while Rocky’s approach could make scary changes much less scary. That’s nerd catnip.
The other camp immediately hit the brakes. One skeptic basically said the launch page gave off “LLM smells” — internet shorthand for marketing copy that feels puffed up or AI-generated — and asked the brutal question: if the intro feels sloppy, why trust the code? Ouch. Then came the platform turf war: does Databricks already do this stuff? Is Rocky overselling what makes it special? Even a friendly dbt employee joined in with the classic tech-world compliment sandwich: cool project, love the branching, but are you maybe dunking a little too hard on dbt and Jinja? In other words, Rocky launched a product — and accidentally launched a debate club
Key Points
- •Rocky is presented as an open-source Rust-based control plane for warehouse pipelines that works alongside platforms such as Databricks and Snowflake.
- •The article highlights features including branching, replay, column-level lineage, compile-time data-contract enforcement, and per-model cost attribution.
- •Rocky provides a local DuckDB-based playground with install scripts and reproducible proof-of-concept demos for schema drift recovery, lineage, branches, and AI-assisted model generation.
- •The project includes multiple subprojects: a Rust CLI engine, a Python Dagster integration, a TypeScript VS Code extension, and playground examples for testing and benchmarking.
- •The repository includes build, test, lint, release, documentation, contribution, sponsorship, and licensing information, with artifacts released independently and the code licensed under Apache 2.0.