Twelvelabs

TwelveLabs is a video intelligence platform (Web, API) for enterprises and developers who need semantic search and deep analysis across large video libraries.

Key capabilities include:

• Semantic video search with natural language queries, returning frame-accurate timestamps across visuals, speech, on-screen text, and audio.
• Video-to-text analysis that generates structured outputs such as scene descriptions, narrative-aware summaries, and other promptable text formats.
• Multimodal embeddings via an Embed API to power semantic and hybrid search, recommendations, and downstream retrieval workflows.
• Workflow integrations and scale features including Frame.io integration, batch processing, and folder-level indexing.

Pricing: Freemium, Free, Developer, and Enterprise tiers with pay-as-you-go usage.

Notable for its proprietary video foundation models, Marengo (semantic search) and Pegasus (video-language understanding), plus the Jockey agent framework for coordinating complex video intelligence workflows. The company has also published research at major computer vision conferences and received investment from Snowflake Ventures.