“They do not reliably capture what a user was shown or told.” This adds to the case for middleware providers like Vapi, LiveKit, and Layercode. If you’re build

“They do not reliably capture what a user was shown or told.” This adds to the case for middleware providers like Vapi, LiveKit, and Layercode. If you’re building a voice AI application using one of these SST -> LLM -> TTS providers there will be definitive logs to capture what a user was told.