Jockey by TwelveLabs .

Jockey by TwelveLabs

Jockey by TwelveLabs is the agent layer for video understanding, transforming raw footage into queryable, structured intelligence. Built on advanced multimodal AI models, Jockey enables marketing teams, creators, and developers to search, analyze, and extract meaningful insights from video libraries using natural language—no manual tagging or folder diving required.

Product Highlights

• Natural-Language Video Search: Describe any moment in plain language and Jockey finds the exact clip with frame-accurate timestamps, not just filenames or folders.
• Automatic Ad Tagging & Hook Detection: Instantly label ads by hook, talent, format, and tone while identifying which opening seconds actually capture audience attention.
• Multimodal AI Architecture: Powered by Marengo for semantic indexing and Pegasus for rich description, delivering embeddings across video, image, audio, and text.
• Seamless Assistant Integration: Connect directly to Claude (with ChatGPT coming soon) to query your video library from the chat interface you already use.
• Structured Data Extraction: Define custom JSON schemas to receive timestamped, machine-readable metadata for automated workflows and deep-linking.

Use Cases

• Marketing Asset Management: Marketing teams can instantly search entire creative libraries to find every ad featuring a specific spokesperson, logo, or visual hook—enabling faster campaign analysis and performance optimization.
• Creator Content Workflows: AI-native creators can talk to their footage archives to locate quotes, generate highlight reels, and produce edit-ready clips without scrubbing through hours of raw material.
• Developer Video Applications: Developers can embed powerful video understanding into their products via API, shipping search bars and retrieval systems that comprehend actual content rather than just metadata.

Target Audience

Jockey serves marketing teams managing large creative libraries, content creators seeking efficient archive navigation, and developers building video-native applications who need production-ready AI infrastructure for multimodal search and analysis.