WrenAI: An Open-Source Semantic Layer That Keeps AI Agents From Guessing at Your Business Logic

Point an AI agent at your company database and it writes SQL that looks right and gives the wrong answer, because it doesn't know status 4 means refunded. WrenAI keeps that business meaning in plain files your team reviews in git, and any coding agent can read them. Queries get expanded and checked before they touch the warehouse. 15.8k stars, Apache-2.0. #TextToSQL #SemanticLayer #ContextEngineering #DataEngineering

Confirmed accurate:

• The project positions itself as a governed text-to-SQL layer for AI agents, turning natural-language questions into trusted SQL, dashboards, and charts, built specifically because business logic like definitions, enums, and approved joins lives outside the database and agents keep getting it wrong — this matches the post's core claim about status codes and hidden business meaning almost exactly. GitHub
• Business meaning, approved definitions, and proven examples are captured as reviewable, version-controlled context rather than being buried in prompts — matches the "plain files your team reviews in git" claim. GitHub
• The core, SDK, and skills are open-sourced under the Apache-2.0 license — confirmed. One caveat: a third-party review notes the core engine is Apache 2.0 while some other components are covered under AGPL-3.0, so "Apache-2.0" is accurate for the core but slightly oversimplifies the full repo's licensing if someone assumes everything in it is Apache-2.0. GitHubSpheron
• The system includes dry-plan validation, structured errors with hints, and value profiling before queries reach the warehouse — supports the "queries get expanded and checked before they touch the warehouse" claim. GitHub
• It's designed to be agent-driven, installable via CLI with a discovery stub so any AI client can read the context — supports "any coding agent can read them." GitHub

Star count — close but slightly off:

• One independent source from roughly a week ago put it at 16.6k GitHub stars and 1.9k forks. The post says 15.8k. Star counts change daily, so this is plausible as a slightly earlier snapshot rather than a fabricated number — not a red flag, just likely a stale figure by a few days/weeks. Spheron

Overall, this is an accurate, well-grounded description of a real, actively maintained open-source tool. No fabricated claims found.