TwelveLabs provides a video intelligence platform and API that uses multimodal AI to search, analyze, and embed insights from video content. Their technology enables users to pinpoint exact moments, generate text summaries, and create vector embeddings across large video libraries. This unlocks deeper understanding and automation capabilities for workflows in media, advertising, and security sectors.
Funding
$107.1M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
+6Founders
Product
Problem
Manually tagging videos is time-consuming and unscalable, while transcripts alone miss critical visual and auditory elements. Existing object-level tags often lack the contextual understanding needed to derive real value from video content. This makes it difficult to efficiently search, analyze, and extract insights from large video libraries.
Solution
Twelve Labs offers a video intelligence platform that utilizes state-of-the-art video foundation models to generate rich, multimodal embeddings. These embeddings enable users to perform precise scene searches using natural language, generate contextual summaries, and unlock deeper insights from video content. The platform's AI models, Marengo and Pegasus, combine temporal and spatial reasoning to understand video in a human-like way, considering visual, audio, and textual elements. This allows for fast, precise, and context-aware search results across speech, text, audio, and visuals. The platform offers APIs for search, analysis (formerly generate), and embedding, allowing developers to build intelligent video applications.
Target Audience
The primary target audience includes businesses in media and entertainment, advertising, automotive, government, and security sectors that manage large video libraries and need advanced video understanding capabilities. This also includes developers looking to build intelligent video applications.
Features
- **Search API:** Enables natural language search across video libraries, identifying specific scenes and moments.
- **Analyze API:** Generates accurate and insightful text about videos through prompting, including summaries, shot lists, and title suggestions.
- **Embed API:** Creates multimodal embeddings from video, audio, text, and image files.
- **Marengo Model:** A powerful encoder model that provides world-class accuracy in video understanding.
- **Pegasus Model:** A native video-language model that enables temporal and spatial reasoning.
- **Customizable Models:** Models can be fine-tuned with user data to specialize in specific content and domains.
- **Flexible Deployment:** Deployable on any cloud, private cloud, or on-premise infrastructure.
- **Scalable Infrastructure:** Designed to handle large video libraries, including petabytes of data.