← Back to live feed1 story
Fish Audio has launched the Drama 3 open-weight text-to-speech model, which provides word-level control and plain English instructions for emotional delivery. The model is available via its app and API, supporting multi-speaker dialogue and specific acoustic attributes.
Sign in to suggest
Key sources
- SOURCE@fishaudio“$52M seed round. 15M+ users.”x.com
- SUPPORT@sgl_project“89ms median time to first audio on a single H200 (single stream)”x.com
- SUPPORT@sgl_project“delivers 72% higher throughput and 37% lower median TTFA at 256 concurrent streams”x.com
- SUPPORT@itsolelehmann“you can just direct the attitude behind the words in plain english”x.com
- SUPPORT@chetaslua“changed one word , {rest of the take stayed the same / it re-read the line}”x.com
- SUPPORT@rohanpaul_ai“Directions go inside square brackets in plain language, such as “barely holding it together,””x.com
- SOURCEmarketbrief.now
- SUPPORT@artificialanlys“Pronunciation Robustness: Drama 3 Preview scores 93.2%, ranking #1 ahead of Eleven v4 at 91.7%”x.com