Videos

Upstart's voice agent for loan verification calls achieved a 92% human-LLM judge agreement rate on compliance by...

Model leaderboards are basically useless for picking voice agent components, because they score individual models on...

Transcripts alone are dangerously insufficient for debugging voice agents because they miss audio-specific failures...

EliseAI's cascaded voice agent architecture uses a streaming speculative transcriber (layering fast Flux on accurate...

Voice agents are moving from demos to production at scale, but reliability remains the top blocker, with even a 1%...

Full-duplex speech-to-speech models solve the walkie-talkie problem of turn-taking, but they pay a fundamental...

Google DeepMind is building a single native speech-to-speech model for Gemini that handles streaming translation,...

Rime Labs' voice models handle numbers, dollar amounts, and IDs accurately, solving a common weakness in voice...