Datasets
Video

Long-Form Narrative

Scripted fiction told over feature runtime or full episodic seasons, with character continuity, performance arcs, and plot structure built across the full duration of each piece. Episodes and films are kept whole rather than chopped into short clips.

Hours9.8K+
Genres10+
Countries30+

Training Use Cases

Long-context video understanding and full-runtime reasoning
Long-form coherent video generation
Character identity persistence across long timelines
Scene boundary and narrative structure detection
Key Highlights
10+ genres covered including drama, comedy, action, thriller, horror, sci-fi, romance, and mystery
30+ countries of origin spanning studio, independent, public broadcast, and streaming productions
Feature films, episodic series, miniseries, and limited runs all represented in the set
Live-action and animated long-form fiction available side-by-side at full runtime

Metadata Fields

durationLength of clip in HH:MM:SS
resolutionPixel dimensions (e.g., 1920x1080, 3840x2160)
frame_rateFrames per second (e.g., 24, 30, 60, 120)
contains_audioWhether the clip carries an audio track (boolean)
primary_categoryDominant content category assigned to a video
styledrama | comedy | action | thriller | horror | sci_fi | romance | mystery | historical | family
formatfeature_film | episodic_series | miniseries | limited_run
runtimeTotal runtime of the piece (HH:MM:SS)
production_originCountry of origin (ISO 3166 code)
live_action_or_animatedlive_action | animated