Crypto
Home›Crypto›Market Structure›5.6 billion TikTok video metadata rows posted on Huggi…
5.6 billion TikTok video metadata rows posted on Hugging Face
The 460 GB dataset spans July 2014 through October 2026 and includes engagement metrics like views, likes, and downloads, plus TikTok Shop product and seller IDs.
A developer posting under the name hashfunction has published metadata for about 5.6 billion public TikTok videos on Hugging Face for free download, according to Decrypt. The dataset, published under the datasocial account, covers videos from July 2014 through October 2026.
The release contains metadata rather than video footage, with each monthly file stored as Parquet and totaling 460 GB across the full set. The fields include captions, hashtags, sound IDs, on-screen text, TikTok Shop product and seller IDs, and engagement counts such as views, likes, comments, shares, saves, and downloads.
Decrypt reports that some columns reflect TikTok labels, including whether a clip is flagged as AI-generated and whether TikTok keeps it off the For You page. Hugging Face indicates the dataset had recorded 1,181 downloads so far.
Decrypt also notes the dataset is not gated, meaning anyone can pull it. The outlet says the metadata availability could be valuable to AI developers building models for what goes viral or sells on TikTok Shop.