A method for generating a target video, an apparatus, a server, and a medium are provided. The method includes: obtaining live broadcast stream data, wherein the live broadcast stream data comprises at least one among voice data and live broadcast interaction data, and video data; performing processing on the live broadcast stream data, and generating at least one among a corresponding voice metric value and interaction metric value, and a corresponding video metric value according to a target object included in a processing result; generating an overall metric value for the live broadcast stream data according to the generated metric values; and in response to determining that the comprehensive metric value for the live broadcast stream data satisfies a preset condition, generating a target video on the basis of the live broadcast stream data.