December 2025 Summaries

Vapi introduces a flexible voice service designed for high-volume, cost-sensitive use cases, allowing users to choose from multiple voice providers like ElevenLabs, PlayHT, and Deepgram without needing to rewrite integrations. As companies scale from a few hundred to hundreds of thousands of calls, the cost of premium voices can become substantial, prompting the need for alternatives in scenarios such as appointment reminders or status updates where speed and cost are prioritized over voice naturalness. Vapi Voices Beta offers a solution with a cost of $0.0025 per minute and optimized latency for responsive conversations, enabling teams to improve unit economics by routing less critical calls to these new voice models while maintaining premium voices for more complex interactions. The release aims to gather real-world data on use cases, latency performance, and cost-effectiveness, with feedback from beta users directly shaping its production release.

Voice agent developers often face challenges when changes to prompts or tools result in silent regressions that aren't immediately detectable, impacting performance metrics over time. To address this, a new testing method called "Evals" has been introduced, which allows developers to write tests for voice agents similar to unit tests for code. Evals involve defining JSON conversations with specific criteria for each assistant interaction, enabling developers to ensure that the assistant performs expected actions, such as calling the right tools or asking necessary questions. This method supports different judging strategies, including exact matching for deterministic behaviors and LLM-as-judge for subjective evaluations like tone and policy adherence, providing a comprehensive testing framework that aligns with continuous integration and deployment pipelines. By turning past production issues into tests, Evals help prevent future regressions, making voice agents more reliable and consistent as they evolve. The approach empowers developers with confidence in their deployments, offering a structured way to maintain and improve the performance of voice agents over time.