Extracts descriptive keywords and tags from visual scenes or video frames for indexing, search, and content summarization.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.