"Someone Open-Sourced the Exercise Database Every Fitness App Is Quietly Built On

the dataset exists, is genuinely comprehensive, and the tooling (SQL generators, API scaffolding, HTML viewers) is legitimate and useful.

 the "shouldn't be free" and "$200/month API" framing implies unrestricted commercial value, but the actual media assets aren't freely licensed for reuse — only the code/metadata structure is MIT. The "bilingual" claim undersells the real 10-language instruction set. This is a case where the headline hook overstates what you actually walk away able to legally do with the most valuable part of the dataset.