“The first model developed by ElevenLabs was a text-to-speech model capable of understanding context to generate appropriate emotion and intonation.”