This connects Claude to Twelve Labs' video intelligence API, letting you index, search, and analyze video content using their Marengo embedding model and Pegasus reasoning engine. You can run natural language searches across video libraries to find specific scenes, actions, or dialogue, generate video embeddings for similarity matching, segment content by scene breaks, and extract timestamped insights from footage. The server exposes their core API operations through streamable HTTP and SSE transports. Reach for this when you're building agents that need to understand video at scale, whether that's searching archives, creating highlight reels from rough cuts, or analyzing hours of footage without manual tagging. Twelve Labs claims 60x real-time indexing speed and supports 47 languages.
claude mcp add --transport http io.twelvelabs-twelvelabs-mcp-server https://mcp.twelvelabs.io