Aug 10, 2026: AI_
Snowflake Cortex AI_MULTI_EMBED is now generally available for semantic video search with support for large video files up to 6 GB using the TwelveLabs Marengo Embed 3.0 model, enabling organizations to index and search production-grade video content directly in Snowflake. This makes it easier to retrieve relevant scenes, quotes, actions, and brands from large video libraries using multimodal embeddings instead of relying only on filenames, tags, or manual review.
Key use cases include:
- Brand suitability and contextual advertising: Evaluate whether video segments align with brand guidelines, safety thresholds, or campaign requirements, and match ads or promotions to relevant scenes, topics, or moments.
- Scene analysis and retrieval: Search large video libraries for relevant scenes, actions, and visual concepts, and analyze how scenes evolve across content.
- Spoken-moment discovery: Find quotes, dialogue, and transcript-aligned moments across long-form video content.
- Brand and product search: Identify where brands or products appear or are mentioned across media collections.
- Content indexing at scale: Build richer media indexes for recommendation, moderation, analytics, and downstream AI applications.
For video inputs, AI_MULTI_EMBED uses the twelvelabs-marengo-embed-3-0 model to generate
embeddings across visual, audio, and transcription signals. As part of a broader workflow, teams can
use these embeddings and derived labels to power downstream ML models, combine video-derived signals
with subscriber or other first-party data, and enable agents to reason over multimodal context and
provide recommendations, all within Snowflake’s governed platform.
For more information, see AI_MULTI_EMBED and Cortex AI Functions: Multimodal.