AssemblyAI Launches Universal-2 Speech API

AssemblyAI · x · 2026-07-15

AssemblyAI announced that its Universal-2 speech model is now live on the Playground. Users can interact with the model, test response speeds, and then call the API using the same request.

Regarding billing, the API has no minimum spend threshold and charges entirely based on the actual duration of audio sent. This means that for massive volumes of short commands (e.g., thousands of 5-second voice prompts), the system won't round up individual durations to charge extra, making it highly friendly for developers handling high-concurrency, short audio clips.

Original post →

More from Apps

Apps channel →