datasocial/tiktok-5.6B-videos

This is a large-scale, public dataset offering metadata for over 5.6 billion TikTok videos from 2014 to 2026. It includes captions, hashtags, sounds, shop links, engagement statistics, and more. The dataset is available in Parquet format, supports SQL querying, and includes access to additional resources like creator profiles, daily histories, scraper source code, and a public ClickHouse database for real-time queries.