Dataset Sample Hub

Narrative Content

Scripted Entertainment 121 hrs sample 31,773 hrs catalog A corpus of high-quality feature films, short films, and TV series across major global languages. All content is studio-grade and contains persistent characters and storylines. Genres include drama, crime, animation, comedy, romance, history, and action. Browse sampler

Unscripted Content

Reality, talk, travel, and lifestyle 100 hrs sample 40,249 hrs catalog Unscripted, documentary-style programming: reality and competition formats, talk and interview shows, travel and lifestyle series, and live event coverage. Natural speech, real locations, and unstaged interaction. Browse sampler

Sports with Play-by-Play

Grounded play annotations 100 hrs sample 1,709 hrs catalog Sport-aware analysis that re-segments game footage into individual plays and grounds them in real game data. Frames are analyzed temporally to detect plays. Results are verified visually to reduce hallucinations and can be reconciled against official play-by-play records (ESPN, college athletics sites, and more). Browse sampler

Professional and Collegiate Sports

Sports broadcasts, profiles, and documentaries 146 hrs sample 120,474 hrs catalog Full-length collegiate and professional game broadcasts, player profiles, and sporting documentaries. Content spans basketball, soccer, table tennis, volleyball, weightlifting, hockey, track & field, rugby, MMA, adventure sports, outdoor sports, eSports, poker, and many more. Browse sampler

Animation and Anime

Classic anime and children's cartoons 99 hrs sample 2,681 hrs catalog The dataset includes classic Japanese anime shows and animated children’s shows and cartoons. Browse sampler

Automotive

TV series, motorsports, and product showcases 102 hrs sample 3,139 hrs catalog A comprehensive automotive video collection featuring award-winning automotive TV series that cover design, engineering, and restoration. Collection further includes product showcases, motorsports, advertisements, and behind-the-scenes interviews. Browse sampler

Science & Technology

Premium science, space, and engineering docs 105 hrs sample 1,016 hrs catalog This dataset contains premium documentaries and factual entertainment spanning science, space, engineering, climate, and nature. It features high-quality non-fiction storytelling, combining immersive real-world footage, expert interviews, and studio-produced TV series. Browse sampler

Premium Wildlife & Natural History

Blue-chip nature and animal behavior 101 hrs sample 4,158 hrs catalog Blue-chip wildlife and natural-history programming: animal behavior in the wild, sweeping landscapes, and macro, aerial, and slow-motion cinematography from leading natural-history producers. Browse sampler

Emerging Subjects and Objects

Subject entries and camera reveals 2.2 hrs sample 833 hrs catalog Clips built around dynamic subject entries and camera reveals: subjects emerging from forests, water, doorways and around corners or arriving from off-screen, alongside pans, dollies and tracking shots that expose new subjects and objects in the environment. Browse sampler

Video Data Sample Pack

A cross-section of the library 2,225 hrs sample 2,225 hrs catalog Clips from a range of categories — facial expression, high body motion, multiple characters, animation and emerging objects — segmented from full-length video across many genres. Browse sampler

High Dynamic Range Video

15-second clips, keyword-searchable 0.6 hrs sample 166 hrs catalog Highly diverse dataset containing 15-second clips (on average) of HDR-enabled video footage. Dataset comes with a user interface searchable by keywords. Browse sampler

Synchronized Multi-Camera

Aligned event groups 169 hrs sample 1,900 hrs catalog Given an arbitrary archive of footage from a multi-camera shoot, we discovered which files belong to the same real-world event and aligned them on a common timeline. An agent-driven discovery pass proposes candidate groupings from filenames, folder structure, and embedded metadata (EXIF, professional timecode); audio fingerprinting then confirms and precisely time-aligns the members. Browse sampler

Character Tracking

Faces clustered into identities 99 hrs sample 4,211 hrs catalog Identifies who appears on screen and when, across an entire show or series. Faces are detected and tracked within each episode, then clustered across videos into unique characters, with optional actor identification. Browse sampler

Instructional & DIY Content

Scored for instructional fit 100 hrs sample 10,000 hrs catalog Step-by-step instructional and DIY content. Each clip is described, tagged, and scored for instructional fit, with a true instructor–learner dynamic distinguished from a solo presenter. Browse sampler

Vice Raw Footage

2006-present raw archive 5.8 hrs sample 450,000 hrs catalog Direct access to Vice Media's complete raw footage archive, spanning 2006 to the present. This extensive, unedited repository offers unprecedented access to thousands of hours of native B-roll, unscripted field reporting, raw interviews, and global investigative footage — providing AI and computer vision teams with a rich, real-world dataset for multimodal pre-training, raw video processing, and scene analysis. Browse sampler

Curiosity Raw Footage

Premium unedited B-roll from Curiosity Stream 100 hrs sample 300,000 hrs catalog Premium raw footage and unedited B-roll from Curiosity Stream's (NASDAQ: CURI) original award-winning films and series. Featuring content viewed globally across major platforms — including Curiosity Stream, Netflix, and Amazon Prime Video — this high-budget factual entertainment archive provides direct access to pristine, unscripted field footage across science, technology, history, and natural history. It delivers an exceptional foundation for multimodal AI pre-training, high-resolution video synthesis, and complex scene understanding. Browse sampler

Stock Footage and Clips

Diverse stock clips and b-roll library 100 hrs sample 10,013 hrs catalog Highly diverse collection of stock clips and b-roll, including animals, bodycam, cars, fails, news, weather, extreme sports, festivals and events, and more. Browse sampler