AI voices you can't tell from humans. Expressive multilingual speech models and developer-first APIs. 👉 https://xitter.cf/t.co/gzh4M4drSq @FishCreativeHQ
San Francisco, CA
Joined April 2024
- Tweets3.8K
- Following31
- Followers36.7K
- Likes844
Fish Audio retweeted
Fish Audio's Drama 3 Preview is the new #1 open-weights Text to Speech model on the Artificial Analysis Text to Speech leaderboard. The model also leads all models in our Pronunciation Robustness leaderboard, ahead of Eleven v4 and Eleven v4 Turbo
Drama 3 Preview is the new open weights Text to Speech model from @FishAudio, built for multi-character scenes and long-form content with voice cloning and natural-language control. Weights are available on Hugging Face under the Fish Audio Community License.
Key takeaways:
➤ Pronunciation Robustness: Drama 3 Preview scores 93.2%, ranking #1 ahead of Eleven v4 at 91.7% and Eleven v4 Turbo at 90.1%. It ranks #1 in Standalone Terms and Preserving Exact Sequences and #2 in Expanding Shorthand.
➤ Provider Voice: Drama 3 Preview has an Elo of 1,243 (+16/-16) across 1,857 appearances, ranking #10, a 101 point gain on the previous Fish Audio S2.1 Pro at 1,142.
➤ Controlled Voice: Drama 3 Preview ranks #14 with an Elo of 1,058 (+15/-15) across 1,804 appearances.
➤ Open weights: Drama 3 Preview is the #1 open weights model on all three of our Text to Speech leaderboards. On Pronunciation Robustness it leads the next open weights model, Mistral's Voxtral TTS at 78.1%, by 15 points, and in the Provider Voice Arena it is ahead of Breeze TTS 2 at 1,221. Drama 3 is a 35B parameter mixture-of-experts model (3B active) supporting 83 languages, available on Hugging Face.
➤ Pricing: Drama 3 Preview is available via the Fish Audio API at $15/1M characters, the same as S2.1 Pro, compared to Eleven v4 at $80/1M characters and Eleven v4 Turbo at $40/1M characters.
See more details and listen to samples below 🧵
Fish Audio retweeted
We'll give you an iPhone Duo for helping us find our next hire.
We're the fastest-growing voice AI company in the world, and we're only as good as who we hire next.
Today 8M+ people build on our models. We're still a lean team building towards $100M ARR. And to scale faster, we're hiring new dedicated roles:
Fish Audio retweeted
"$99 al mes". Eso cuesta el plan Pro de ElevenLabs para generar voz con IA.
Esto lo hace gratis en tu propia máquina. Y acaba de superar a los modelos cerrados.
Se llama Fish Audio S2 y ya tiene 30.000 estrellas en GitHub.
Entrenado con 10 millones de horas de audio en más de 80 idiomas, hace lo que hasta ahora solo veĂas en herramientas de pago:
→ clona tu voz con una muestra de 10 a 30 segundos
→ controla la emoción con etiquetas como [whisper], [excited] o [angry] en cualquier punto del texto
→ genera diálogos con varios hablantes en una sola pasada
→ arranca el audio en ~100 ms, listo para apps en tiempo real
Y los benchmarks lo confirman:
en Seed-TTS Eval registra el menor error de todos los modelos evaluados, por delante de Qwen3-TTS, MiniMax y Seed-TTS.
Funciona en español, francés, inglés y decenas de idiomas más. No necesita fonemas ni preprocesado por idioma.
Open source, con el modelo de 4B parámetros disponible en HuggingFace.
Repo aquà ⬇️
Fish Audio retweeted
Fish Audio has 2M+ voice models built by the community. That comes with a responsibility we take seriously. If someone uploaded your voice without permission, we will take it down.
We built a formal DMCA process for voice actors and rights holders to report and remove unauthorized voice models.
To file a request, you'll need:
1. A valid government-issued ID (or authorization letter if filing on behalf of someone)
2. Documentation showing the voice is yours (professional credits, contracts, press coverage, or verified social media announcements)
Submit directly from the model's page or through our support cases dashboard. We review every request and follow up.
Read the full guide → fish.audio/blog/dmca-request…
Fish Audio retweeted
Dario after realizing the code Claude wrote is now on GitHub with 100k stars
voice by @FishAudio
Fish Audio retweeted
we made a pr @joshavant 🥺
github.com/openclaw/openclaw…
Fish Audio retweeted
$0 → $10M ARR in 12 months.
Open source.
No sales team.
No paid ads.
Today:
• 5M users
• 1.5M+ MAU
• 2M public UGC voices (largest voice library in the world)
• ~50% revenue from enterprise
Here's the playbook behind @FishAudio🧵
Fish Audio retweeted
Yesterday, we wrapped up a hands-on workshop at the AWS GenAI LOFT in San Francisco!
We’re thrilled that over 40 developers joined us, each building their own multimodal AI agents from scratch using @TenFramework. The TEN team was there to assist with tech questions, sparking lots of creative ideas. Some participants plan to develop a vertical voice agent for tumor boards, while others considered creating a buddy agent for their children. Six participants completed our challenges, and we’re glad to see how much they enjoyed using @TenFramework and TEN Agent!
Our friends at @FishAudio were also there, delivering a great talk on TTS and voice cloning, and they announced their latest products during the event!
A big thank you to @AgoraIO and @aicampai for supporting this fantastic event!
Stay tuned, and we hope to see you at the next TEN Meetup/Workshop!
