The present application discloses techniques for interaction during live video streaming. The techniques comprise obtaining and playing streamer-end video data, and user-end video data captured by a user terminal in real time; monitoring and recognizing whether the streamer-end video data comprise a preset voice instruction; determining whether the user-end video data comprises a target audio or a target video when the streamer-end video data comprises the preset voice instruction; and displaying a video special effect corresponding to the preset voice instruction in a user video when the user-end video data comprise the target audio or the target video. By means of the present application, a video special effect can be played for a user video according to a result of interaction between a streamer and a user, which enriches the way of interaction presentation and enhances the sense of participation in interaction.