Fish Audio
Fish Audio is a deep learning-based AI speech synthesis platform that provides text-to-speech (TTS) services and voice cloning functionality, including zero-shot voice cloning technology capable of replicating a target voice from a small number of audio samples. The platform targets developers and content creators, offering API access and pre-trained voice models. Users can share and utilize a variety of voice models on the platform, with applications spanning dubbing, virtual assistants, and content creation.
Timeline (last 90 days)
Fish Audio released the speech model Drama 3, which ranks first among open-weight models on both of Artificial Analysis's speech arena leaderboards, ranks first among all models in pronunciation robustness tests, and scored 1278 ELO when tested on text-to-speech using voices provided by the service provider.
Fish Audio released Drama3, an open-weight text-to-speech model that supports natural-language direction of vocal performance, switching tone within a single sentence, and controlling the expression of individual words; it is already live on the web, mobile app and API, and users can download the weights to run it themselves.
Drama 3's access methods cover web, mobile, and API, and under the Fish Audio community license it supports open-weight self-hosting.
Fish Audio released the Drama 3 model on Product Hunt, focusing on directability (natural-language control of emotion, pacing, and pauses).
Drama 3 is released with open weights, accompanied by a revenue-capped community license, and commercial use has a revenue threshold.
Drama 3's API price is $10 per million characters, rising to $50 after November 12, 2025.
Fish Audio released Drama 3 in October 2026, a 35-billion-parameter speech generation model, with weights published on Hugging Face.
Drama 3's interactive editing feature is currently invitation-only and not yet launched.
Drama 3 allows natural-language performance instructions to be embedded directly in the script text (such as bracketed emotion and pacing annotations, and <whisper> and <emphasis> tags) to control emotion, tone and pauses.
Fish Audio's text-to-speech model Drama 3 has officially launched on AI Gateway and is currently in preview.
3 more timeline events
All Facts (10)
Fish Audio是一款基于深度学习的文本转语音（TTS）服务，支持多种音色克隆和自然语音合成
90%UnverifiedFish Audio采用的技术路线支持零样本（Zero-shot）语音克隆，即只需少量参考音频即可复制特定音色
85%UnverifiedFish Audio released the speech model Drama 3, which ranks first among open-weight models on both of Artificial Analysis's speech arena leaderboards, ranks first among all models in pronunciation robustness tests, and scored 1278 ELO when tested on text-to-speech using voices provided by the service provider.
50%UnverifiedDrama 3's access methods cover web, mobile, and API, and under the Fish Audio community license it supports open-weight self-hosting.
50%UnverifiedDrama 3 is released with open weights, accompanied by a revenue-capped community license, and commercial use has a revenue threshold.
50%UnverifiedDrama 3 allows natural-language performance instructions to be embedded directly in the script text (such as bracketed emotion and pacing annotations, and <whisper> and <emphasis> tags) to control emotion, tone and pauses.
50%UnverifiedThe core reason for choosing S2.1 Pro was its ability to maintain the same voice when switching between different languages such as English and Hindi.
50%UnverifiedFish Audio released Drama3, an open-weight text-to-speech model that supports natural-language direction of vocal performance, switching tone within a single sentence, and controlling the expression of individual words; it is already live on the web, mobile app and API, and users can download the weights to run it themselves.
50%UnverifiedFish Audio released the Drama 3 model on Product Hunt, focusing on directability (natural-language control of emotion, pacing, and pauses).
50%UnverifiedFish Audio released Drama 3 in October 2026, a 35-billion-parameter speech generation model, with weights published on Hugging Face.
50%