Tolan builds voice-first AI with GPT-5.1 (openai.com)

Tolan has launched a groundbreaking voice-first AI companion powered by GPT-5.1, designed to facilitate natural, fluid conversations that evolve over time. This innovative app, developed by Portola, prioritizes low latency, contextual accuracy, and stable character personalities, allowing users to interact with personalized, animated avatars. CEO Quinten Farmer emphasizes the challenges of voice AI compared to text-based systems, noting that maintaining engaging dialogue in real-time requires sophisticated context management and interaction design. The significance of Tolan lies in its advanced architecture that addresses the unique demands of voice conversations. By employing a real-time context reconstruction approach and leveraging a high-speed vector database for memory recall, Tolan dynamically adapts to shifting topics without losing continuity. This innovative methodology, along with improvements in memory and character design, has resulted in noticeable enhancements in user experience, including a 30% reduction in memory recall misses and a 20% increase in daily user retention. With a user base of over 200,000 and a stellar 4.8-star rating, Tolan is positioned to redefine voice interaction by aiming for a truly multimodal AI experience that integrates voice, vision, and contextual awareness.