Hume's Empathic Voice Interface (EVI) is a comprehensive API for building voice AI applications with emotional intelligence. This voice-to-voice API processes live audio to understand both verbal content and emotional expressions, and generates responses that match the context and emotional tone of the conversation. EVI excels in natural conversation flow, featuring precise turn-taking and interruptibility. The system's expressive text-to-speech produces nuanced, emotionally appropriate vocal responses that sound remarkably human-like. Developers can tailor EVI's behavior through prompts, integrate it with different language models, and extend its functionality via function calling to allow the system to use tools. With its focus on empathic AI and natural interaction, EVI enables creating more engaging and effective voice-based applications in fields ranging from customer service to healthcare. For instance, in sales, EVI can handle scheduling requests with natural conversation, using calendar tool while adjusting its tone to match the user's urgency or preferences. EVI can act as an emotionally intelligent AI customer support agent, detecting customer frustration through voice analysis and responding with empathy while efficiently resolving issues using database lookup and order management tools. In education, EVI can function as an interactive tutor, adapting its teaching style based on the student's emotional responses and using tools to access educational content or update progress.