Skip to content
C-Kernel-Engine docs Antshiv Open Source

C-KERNEL-ENGINE · GUIDE

Generate speech

Which text-to-speech option should I try?

Kokoro: start with a short sentence and the voice and model bundle specified in the Kokoro guide. CKE has generated audible-duration WAV output and measured synthesis time on a particular CPU. Listen for missing words, clipped endings and unexpected pauses.

Ming3 0.5B: experimental text-to-WAV synthesis includes a sentence-chunked 100-word trial. Start with the supplied example before attempting a long passage. Speed, listening quality and a hosted endpoint remain open items. Ming3 instructions

Qwen3-TTS: CKE has a pinned session and waveform evidence. Use the recorded session to establish what ran before choosing this experimental path. The Whisper audio command takes speech as input; synthesis uses its own model-specific workflow.