// the find
hasaneyldrm/exercises-dataset
1,324-exercise fitness dataset — animation GIFs, 180×180 thumbnails, muscle-group & equipment data, and step-by-step instructions in 6 languages. The exercise data layer behind the LogPress app.
A static dataset of 1,324 fitness exercises with animation GIFs, 180×180 thumbnails, structured muscle/equipment metadata, and step-by-step instructions in 10 languages. It's the data layer behind the LogPress workout app, extracted and packaged for anyone building a fitness application who doesn't want to compile this from scratch.
The multilingual instructions are genuinely useful — both as a flat string and as a pre-split step array, so you don't have to parse them yourself. The JSON Schema (Draft 2020-12) is included, which means you can validate the dataset or your own additions with any standard validator rather than guessing field types. The setup.html wizard that generates ready-to-run INSERT statements for four different databases directly in the browser is a practical touch that removes real friction for people who just want to get it loaded. Coverage across equipment types is solid — 325 bodyweight-only exercises means it's usable for at-home apps without filtering out half the content.
The media is © Gym visual and redistributed 'with permission' under terms that require you to obtain your own license from Gym visual before reusing it — so you can clone this repo and browse it, but you can't actually ship it in a product without separate negotiation. The dataset is frozen at 1,324 entries with no clear contribution path or versioning strategy, so if an exercise is missing or wrong, there's no obvious way to fix it. The `category` and `body_part` fields carry identical values on every record, which suggests a schema collapse that was never cleaned up. There's no difficulty, intensity, or volume guidance per exercise — all exercises are structurally equal, so building a progressive program on top of this requires significant domain logic you'll have to write yourself.