This content benchmarks the capability of frontier Automatic Speech Recognition (ASR) systems and voice agents to process code-switched speech from bilingual customers. It evaluates their performance in understanding and responding to such complex linguistic inputs.
This post introduces the Nova Sonic Test Harness, an open-source framework designed for scalable evaluation and rapid iteration of Amazon Nova Sonic voice agents. It automates multi-turn conversations, uses LLM-as-judge techniques to validate quality, and can detect audio hallucinations without requiring a microphone.
This post demonstrates how to combine Stream's Vision Agents open-source framework with Amazon Bedrock and Amazon Nova 2 Sonic to build production-ready real-time voice agents. It covers the integration mechanics, provides code examples, and explores advanced features like function calling and multilingual support.
This post teaches how to use Amazon Nova Sonic, Amazon Bedrock AgentCore, and Strands BidiAgent to build scalable and maintainable voice agents. It explores popular architectural patterns for voice agents, highlighting trade-offs and best practices for minimizing latency and achieving more intelligent customer interactions.
OpenAI has launched its Realtime API and three new audio models, aiming to make real-time voice AI more accessible and effective. These models are designed to address the previous challenges of cost and complexity in real-time audio AI.
OpenClaw 2026.4.25 has been released, introducing voice agents for natural language understanding and a major plugin overhaul. This update allows local AI model execution, bypassing cloud dependencies, and significantly enhances AI development for users and developers alike.
The AI landscape is buzzing with exciting developments, including breakthroughs in Computer-Aided Design (CAD) tasks and the emergence of low-latency voice agents. These advancements point to a transformative year for artificial intelligence, driving efficient design processes and democratizing voice interfaces.