Show HN: Grapple, a DIY data platform that uses AI the Right Way

Hello again, HN! My name is Andrew, and my cofounder Jack (@jackonawalk) and I are excited to show off a project years in the making – a data platform where the only prerequisite is data.

Most "self-service" data tools aren't really self-service. They assume you've already set up a data warehouse, configured pipelines, defined semantic data models, and usually, the best escape hatch is to explore the underlying data warehouse yourself. That's an intimidating proposition for most people who don't consider themselves data experts.

Grapple abstracts the data architecture and ships with everything you need to start working with the data you already have. It comes with a data warehouse, an ETL for connecting data from existing sources, semantic data models, and a visualization layer to build dashboards and reporting. All you need to bring is your data.

Oh, and the AI thing? We're bullish on safe, secure, and governable use of AI. We use a LLM to help generate calculations and formulas, but it never has access to the underlying data, only the schema. Everything the LLM generates is auditable and editable by the user. We believe this is the most secure way to get the benefits of a LLM while preserving absolute data integrity – no hallucinated data!

You can try it out with just a CSV upload. No credit card or sensitive data required. There's a demo video on our homepage and a gallery of public dashboards here: https://cloudgrapple.com/gallery

Let us know what you think!