Introducing the Sync API

AssemblyAI announces the Sync API, a new hybrid API designed to bridge the gap between real-time and pre-recorded transcription. This API allows users to send audio snippets ranging from 80 milliseconds to 120 seconds and receive transcriptions with a latency of under 300 milliseconds. It supports features like prompting, language steering, keyword boosting, and conversational context, enabling developers to build custom voice agents with end-to-end control over orchestration. The video demonstrates how adding conversational context, such as questions from an agent, significantly improves the accuracy and formatting of transcriptions, leading to a better user experience.