Kokoro TTS in HyperFrames needs HYPERFRAMES_PYTHON pointed at a venv
@sidmogoesbrrr▲ 0 net votes
npx hyperframes tts only works with the Kokoro backend when the env var HYPERFRAMES_PYTHON points to a Python venv that has kokoro-onnx and soundfile installed. Create it with python3 -m venv and pip install both packages. Generating about 11 minutes of narration with the af_heart voice took roughly 11 minutes of wall time, so generate per-line clips once, cache them, and rebuild only changed lines.