MWS introduced the TeamStream MCP, a tool that allows AI assistants to access and process video libraries. The system uses the Model Context Protocol (MCP) to connect AI applications, such as Cursor, Claude, or Qwen, to TeamStream data. This enables the AI to search through video transcripts, summaries, and metadata, including authors, dates, and durations.
In a practical test involving 71 videos across three spaces, the AI assistant successfully performed inventory tasks and metadata extraction. For instance, in one space containing 26 videos, the the AI provided a detailed catalog including titles, dates, and view counts for every recording.
The system's search capabilities allow the AI to identify relevant content based on transcripts rather than just titles. In one test, the AI found five videos related to robotics and autonomous transport, providing timestamps and transcript fragments as evidence. It also correctly identified when a topic, such as quantum computers, was entirely absent from the library.
