December 2025 Summaries

Thymia, founded by neuroscientist Emilia Molimpakis, leverages voice AI technology to analyze speech for mental health assessments, detecting indicators such as depression, anxiety, stress, and other health markers. Utilizing clinically labeled datasets for accuracy, thymia's models can provide insights from just 15 seconds of speech, applicable in both healthcare and safety-critical industries like automotive and airlines. In these sectors, first responders demonstrate a preference for engaging with empathetic voice agents over human peers, leading to longer, more open interactions. The integration of Rime's expressive voices enhances engagement, encouraging users to share personal information while maintaining politeness. Thymia's technology is evolving to offer real-time monitoring and action, detecting stress and inconsistencies in speech to inform safety protocols, thereby creating a responsive and empathetic AI system.

Rime has introduced a new pricing model to simplify the adoption and scaling of its real-time voice API, offering three distinct plans: Starter, Growth, and Enterprise. Each plan is designed for different stages of use, with the Starter plan aimed at prototyping with no overages and $200 in free credits, the Growth plan providing lower unit costs and higher concurrency for sustained workloads, and the Enterprise plan offering customized options for high-scale and compliance-heavy deployments. Existing customer plans remain unchanged unless they opt to upgrade, and the updated model aims to provide transparent, flexible, and predictable costs aligned with real-world workloads. Rime encourages users to explore these plans through their dashboard or contact sales for further assistance.

Rime's speech models have been integrated into Together AI, enhancing the development of AI voice applications with fast, natural, and enterprise-ready capabilities. Together AI, designed for AI-native companies, offers a cloud environment optimized for GPUs, providing low-latency, cost-efficient inference, and customization for always-on production. This integration allows developers to leverage Rime's models within a secure and dedicated environment, ensuring data security and high performance. Rime's models are tailored for high-volume enterprise telephony use cases, offering superior latency and concurrency, leading to significant improvements in call containment and conversion rates. The collaboration between Rime and Together AI provides a seamless deployment experience for enterprise teams, supporting scalable and reliable voice applications with dedicated endpoints and co-located compute capabilities. This combination enables rapid prototyping and deployment of voice agents, with a robust infrastructure that includes dedicated GPU capacity, compliance with industry standards, and a comprehensive developer experience. Existing Together AI customers can easily adopt Rime's models, ensuring an efficient path from prototype to production for voice applications.

Rime is now integrated into Hathora's Models, a model marketplace that offers developers access to advanced ASR, TTS, STS, and LLM models across more than 14 global regions without the need to manage infrastructure or multiple vendors. This collaboration allows developers to easily incorporate Rime's high-quality voice capabilities into their applications using a single API endpoint, enhancing the ease of deploying machine learning and AI workloads. With this integration, developers can effortlessly run Rime's Arcana TTS model in a production setting without dealing with infrastructure setup or container optimization. To celebrate this partnership, Hathora is offering $10 in credits to new users, who can get started by following the provided documentation, with no credit card required.

Goodcall has revolutionized voice agent technology by integrating Rime's advanced speech capabilities, shifting from impressive initial models to delivering conversations that feel genuinely human. As customer expectations evolve towards instant, natural, and human-like interactions, Goodcall's systems now offer superhuman capabilities such as parsing multiple requests, extracting essential details with minimal engineering, and remembering caller preferences. Rime's ultra-low-latency, expressive speech solutions allow these voice agents to match the intelligence behind them, providing a more authentic and emotionally resonant experience. By focusing on conversational flow and fine-tuning responses through A/B testing, Goodcall has successfully created a voice experience that resonates with small and mid-sized businesses, offering enterprise-level AI capabilities that enhance customer satisfaction and set new industry standards.

Rime has launched Dashboard 2.0, a significant update designed to enhance user experience by making it easier to explore voices, generate speech, and utilize everyday tools. The revamped dashboard features an intuitive signup flow, a redesigned home with quick links for generating speech and managing team collaborations, and improved access to support and integration tools. New keyboard shortcuts have been introduced for a more efficient workflow, and updates to real-time monitoring and pronunciation dashboards aim to improve quality assurance. Additionally, the dashboard includes guided demos to showcase various use cases for creating agent voices and real-time interactions, along with a new tool for designing voice prompts. Rime encourages user feedback to continue evolving the platform.