Massive TikTok Video Metadata Collection Released for Free
- A developer named hashfunction released metadata for approximately 5.6 billion public TikTok videos on Hugging Face, spanning from July 2014 to October 2026.
- The dataset includes captions, hashtags, sound IDs, and engagement metrics like views and likes, totaling around 460 GB in Parquet files.
- TikTok’s terms prohibit automated scraping without approval, but the data was allegedly sourced from TikTok’s private mobile API without requiring a login.
- The dataset is free for non-commercial use under a CC BY-NC license, while commercial use and access to creator profiles are managed by datasocial.ai.
- Hugging Face has recorded over a thousand downloads of this dataset so far.
This extensive collection of TikTok metadata offers valuable insights for AI developers interested in viral content prediction and language model training. The release highlights ongoing tensions between data accessibility for research purposes and platform restrictions on automated scraping activities.(Source)
