The Edge voice TTS supported by default in Hermes comes with multiple free voices, synthesizes quickly, is completely free, and performs synthesis in the cloud, so no local speech LLM is needed.