Jockey Agent
AI agent · Workflow / back-office agent · Operations & supply chain by Twelve Labs
The video intelligence agent that sees, reasons, and acts on your video library.
Jockey is TwelveLabs’ video intelligence agent designed for organizations managing large photo and video libraries. It allows users to interact with their footage using natural language to find specific moments, compare creative content, or identify patterns across hours of video. The agent functions by planning its own reasoning steps and can be integrated directly into popular AI assistants like Claude and ChatGPT or accessed via a dedicated Agents API. Jockey is built to solve the "review bottleneck" by automating the search and analysis of video, audio, and on-screen text. It is used by te
- Price
- From 2 USD per 1M tokens
- Languages
- English, Korean
- Model architecture
- TwelveLabs Pegasus and Marengo foundation models
- API support
- Python and Node.js SDKs
- Input types
- Video, audio, images, and documents
- Deployment options
- Cloud, private cloud, or on-premise
Overview
Jockey is TwelveLabs’ video intelligence agent designed for organizations managing large photo and video libraries. It allows users to interact with their footage using natural language to find specific moments, compare creative content, or identify patterns across hours of video. The agent functions by planning its own reasoning steps and can be integrated directly into popular AI assistants like Claude and ChatGPT or accessed via a dedicated Agents API. Jockey is built to solve the "review bottleneck" by automating the search and analysis of video, audio, and on-screen text. It is used by teams in media, advertising, and retail to turn raw footage into actionable insights and structured data.
Video Intelligence Capabilities
- Natural language interaction: Users can ask questions about their video library in plain English to find specific moments or spot patterns.
- Multimodal reasoning: The agent reasons across video, audio, and on-screen text to confirm context and intent.
- Autonomous planning: Jockey plans its own steps to answer complex questions across multiple videos.
- Cited results: The agent provides cited moments and timestamps for its answers to ensure accuracy.
Integrations and Access
- ChatGPT integration: Connects to ChatGPT to search and analyze videos without leaving the conversation.
- Claude integration: Integrates with Claude to make video libraries accessible to the AI assistant.
- Agents API access: Available via the Agents API for developers to build custom agentic workflows in Python or Node.js.
- Flexible ingestion: Supports ingesting videos directly from public social profile URLs or local device uploads.
Use Cases
- Content assembly: Assists in finding matching moments to assemble video cuts, reducing manual hunting time.
- Creative review: Identifies creative patterns, brand inconsistencies, and content gaps in video campaigns.
- Operational intelligence: Analyzes retail CCTV footage to provide insights on operations, security, and customer behavior.
Pricing
- Jockey Responses (Overage)
- $2 per 1M tokens
Specifications
- Model architecture
- TwelveLabs Pegasus and Marengo foundation models
- API support
- Python and Node.js SDKs
- Input types
- Video, audio, images, and documents
- Deployment options
- Cloud, private cloud, or on-premise
Added by AntSpace from public sources (twelvelabs.io).