Cannot Model Conversation Structure
English Language Dominance
Limited Audio Plays on Lower Tiers
Starter: 1/day, Premium: 5/day, Professional: 10/day
Memory Bottleneck in Training
No Explicit Mention of Offline Use
No Mention of CMS Integrations
No Mention of File Format Support Beyond MP3
No Pre-trained Language Model Use
Play Counting Rule
Only new text played is counted as a play. Long pages/articles might be counted as two plays.
Real-Time Generation Delay
RVQ time-to-first-audio scales poorly