Skip to main content
Arka can quickly cache exercise metadata from hasaneyldrm/exercises-dataset for fitness demos, workout planners, local search, and prototype data flows.
Natural language also routes to the same importer:

What is cached

The default import downloads data/exercises.json only and stores normalized metadata under Arka’s cache directory:
Records include exercise name, body part/category, equipment, target muscle, secondary muscles, bilingual instructions, thumbnail path, and GIF path. The upstream JSON is currently larger than 10 MB. If your network is slow, use --timeout to give the import more time instead of retrying through chat.

License and media note

The upstream README describes the dataset as educational / non-commercial only. Images and GIF animations belong to their respective copyright holders. Arka therefore caches metadata by default and does not bundle the dataset or media inside the Arka package. Verify upstream terms before using the data or media in a commercial product.