AssemblyAI offers customized models for different use cases and applications. Select from a below model to see the accuracy on a handful of different applications.
Please select a model from the left to load the demo.
The Default model is generalized across multiple audio types, and pretty good at each. This model is similar to other off-the-shelf, uncustomized models you can get from other providers.
The Phone Call model is optimized for conversational, low quality, compressed audio created by the telephony infrastructure.
The Broadcast Media model is optimized for TV Shows, Conferences, Lectures, and Videos.
The Medical model is optimzed for medical vocabularies and read-speech.