A system, method, and processor-readable storage medium for providing sound effects are disclosed. An example method comprises receiving an audio signal associated with a speech of a user, performing speech recognition on the audio signal to identify one or more recognized words, identifying at least one trigger word among the one or more recognized words, and providing to the user at least one sound effect associated with the at least one trigger word. The speech recognition can be implemented by a trained machine-learning system. The sound effects can be provided to the user and optionally to other users, for example, via a network gaming environment. The sound effects can also be adjusted or selected based on a conversational context, user identification, and/or audio characteristics of the received audio signal. The system for providing sound effects can include a smart speaker, mobile device, game console, television device, and the like.