Blog Post Background :
Okay, so where to begin ? This is the start of a true crime investigation in preparedness pre-production for a several part television documentary series, that looks at current technology that we all use, that affects our rights and freedoms.
Here, I focus upon one standout company called - FISH AUDIO AI - who setup to clone voices without consent and sell them.
Time for the intro video
https://r2.openaudio.com/blogs/s1/launch-video-0603.mp4
Website url : https://fish.audio
So - Whats The Problem ?
🚨 WARNING 🚨
A AI website has stolen and cloned a large number of voices internationally, primarily from gaming backgrounds. I advise voice actors, talent and presenters to take a moment to check this page, and use the search box to search for your name, as well as any character names in video games you have been in:
CLICK THIS LINK --> https://fish.audio/discovery/
It appears most voices appear under their character name for video games rather than the actor’s name. If you find your voice cloned, you can "try' to contact FISH AUDIO to request the immediate removal from their website.
Please share this with friends who may be affected by this breach !
Here's an example of a famous voice :
Example : Meet Elon Musk -> https://fish.audio/m/03397b4c4be74759b72533b663fbd001/
Listen.. does it really sound like Elon ? what if was.. improved and made better. Imagine if this fake voice was used for things by bad actors. All without Elon's consent, yet in his voice (or soundalike).
Fish Audio Development :
So what's the code used for this "ai" ? well you can find it here on Github.
Where you can ALSO find the three developers behind it, who are students :
- Leng Yue - https://github.com/leng-yue
- Stardust minus - https://github.com/Stardust-minus
- Whale Dolphin - https://github.com/Whale-Dolphin
- these aren't their real names, as you can see.. but dear reader please read on, where I shall expose their actual names and where they live in San Francisco.
They develop as a company called "openaudio", see here : https://openaudio.com/blogs/s1
OpenAudio is the Research Lab of Hanabi AI Inc., and Fish Audio is our product platform through which we bring these innovations to the community. We are committed to advancing the field of audio synthesis and look forward to seeing what you create with OpenAudio S1!
So - Who Actually Are Fish Audio ?
Question 1 : How are they funded ?
They are funded by Angel funding, and it's Millions of US Dollars. But who are they ?
Fish Audio closed a financing round right before acceptance into the HF0 incubator. This back-to-back validation catapulted the company’s revenue to $4 million in mere months and lifted its valuation, paving the way for the next strategic raise post-S1 model launch.
website url : https://www.hf0.com/
team behind it : https://www.hf0.com/team
NewYork Times url : https://www.nytimes.com/2023/05/31/magazine/ai-start-up-accelerator-san-francisco.html
linkedin profile : https://www.linkedin.com/company/hf0
Video to watch on this startup : https://www.youtube.com/watch?v=dtNnqqsNWNs
The company founder is : Dave Fontenot, Co-Founder of HF0
At HF0, 10 teams live in a 20,000-square-foot mansion in San Francisco for 12 weeks to build their life's work. During the residency, everything is taken care of for them, healthy meals are provided daily, and founders can focus solely on building their company. The quality of these founders is awe-inspiring, with many teams reaching $2M-$3M+ in ARR by demo day.
[ more to be added here soon ]
Question 2 : Are they known as a business and who are they ?
Well, they are only Fish Audio Ai by name , yet their website points to TWO other named companies, so lets look at them.
Business 1. Hanabi AI Inc
About Hanabi AIHanabi AI Inc. is pioneering the era of the AI Voice Actor—speech that you can direct as easily as video, shaping every inflection, pause, and emotion in real time. Built on our open-source roots, the Fish Audio platform gives filmmakers, streamers, and everyday creators frame-perfect control over how their stories sound.
Businesswire blog say's this :
More business information listed here :
Adress is :
900 Pepper Tree Ln APT 1617 Santa Clara, CA, 95051-5285 United States
this address looks like a residential street in Santa Clara, south of San Francisco. It is ! want to view the house and view the house for sale listing ?
It's 1 bed / 2 bed apartments. It's NOT a big business.
More SEC business company filing is found here :
which states -
- Hanabi Ai Inc., a Stock Corporation, is a business entity located in Santa Clara, CA.
- Officially filed on December 13, 2024.
- This corporation is recognized under the document number 6493519.
- Governed by the California Secretary of State, the corporation maintains an active filing status.
- The principal and mailing address of the corporation is situated at 900 Pepper Tree Ln Apt 1617, Santa Clara, CA 95051, serving as the central hub for its business operations and correspondence.
Note : however, when I serach for this company at https://www.sec.gov/ where all American based companies have to file their business accounts at... I cannot find Hanabi Ai Inc or Fish Audio Ai listed.
hmm.
And if either Fish Audio Ai / Hanabi Ai inc are a SMALL COMPANY, then why aren't they following what the US government sets out clearly here on the gov website :
Business 2 : 39 Ai
So who are 39 Ai ?Website url : https://39-ai.com/
39 AI Inc. is a Bay Area-founded, AI-native startup building more natural and emotionally resonant AI product experiences. We explore more authentic, expressive, and companion-like ways for people and AI to connect.
Our current products include the AI audio product FishAudio, AI chat products, and several new AI product directions still in incubation. The team is flat, direct, and fast-moving. We encourage people to speak up, own ideas, and turn them into shipped work quickly.
The team is young, sincere, relaxed, and creator-minded, with a strong anime, comics, and gaming culture, and deep creator roots. We do not set limits around degrees, compensation, equity incentives, office location, or growth paths. We are looking for people who are deeply excited about AI, products, creation, and new things to explore the next generation of AI products with us.
People :
Our current products include the AI audio product FishAudio, AI chat products, and several new AI product directions still in incubation. The team is flat, direct, and fast-moving. We encourage people to speak up, own ideas, and turn them into shipped work quickly.
The team is young, sincere, relaxed, and creator-minded, with a strong anime, comics, and gaming culture, and deep creator roots. We do not set limits around degrees, compensation, equity incentives, office location, or growth paths. We are looking for people who are deeply excited about AI, products, creation, and new things to explore the next generation of AI products with us.
quote taken from their careers page here : https://39-ai.com/careers
Want to be a Product Marketing Manager for FISH AUDIO Ai and the start-up ? well, there's a job opening for you here :
just email them at : careers@39-ai.com
It's co founder is Shijia Liao , who ALSO created Fish Audio Ai .
They are listed on Linkedin here : https://www.linkedin.com/in/leng-yue
https://www.linkedin.com/posts/rissacao_my-co-founder-shijia-got-into-voice-ai-activity-7458538709135671296-ZO3j
My co-founder, Shijia, got into voice AI almost by accident, and ended up building a TTS model that now challenges companies more than 10x our size. Shijia is fan of VTuber streams: anime-style virtual creators whose performances depend heavily on voice quality. But he kept running into the same problem. Existing TTS systems could sound convincing for a short sentence, then quickly fall apart into a flat, monotone delivery as the dialogue got longer. The entire category seemed capped at around ten seconds of believable speech, and as a power user, it frustrated him. So, he started Fish Speech as an open-source side project to fix that issue. The repo quickly attracted a community of game developers, indie creators, and fellow VTuber fans who used it to dub characters, create content, and experiment with expressive voice generation. By mid-2024, Shijia left NVIDIA to focus on the project full-time. And the open-source community that had formed around Fish Speech became the foundation of Fish Audio’s earliest customer base. Around the time the company started getting real revenue, I met Shijia through mutual friends. My first instinct was to hire him into my own startup. His tenacity got me instead. I jumped ship and joined his. Best friends and partners in crime ever since. Fast forward to early 2026: we released S2 Pro, models that now rank 1 across multiple benchmarks. In a blind production test with more than 5,000 real-user comparisons, S2 achieved a 66% win rate against ElevenLabs. The repo has since crossed 100K GitHub stars. And our ARR reached $13M. And every product decision we have made traces back to one frustration. Shijia could hear when a voice stopped sounding human, so he built a model that doesn't.
https://www.allvoicelab.com/blog/fishing-for-voices-with-fish-audio-from-trump-ai-voice-to-lebron-ai-voice
which states the following :
From viral celebrity impressions to professional AI voice synthesis—unlock limitless creativity with on-demand Text-to-Speech and Voice Cloning.
AI voice technology has become a cultural phenomenon, dominating search trends with phrases like “Trump AI voice,” “LeBron AI voice,” and “Scarlett Johansson AI voice.” At the same time, niche communities are buzzing about the “trapaholics voice generator” and “Fish.audio AI Digimon” samples. Leading the charge is Fish Audio, a browser-based platform that combines Text-to-Speech, AI Voice Cloning, and interactive Dubbing tools into a playful sandbox for creators
Origins of Fish Audio and Shijia Liao’s Vision
Behind Fish Audio is Shijia Liao, whose passion for voice synthesis began in his teens with open-source projects like So-VITS-SVC and GPT-SoVITS. After graduating from the University of Maryland in 2023, he joined NVIDIA to scale Vision Foundation Models across large GPU clusters.
They are listed on Linkedin here : https://www.linkedin.com/in/leng-yue
His partner/girlfriend is Rissa Cao , who also created Fish Audio Ai.
She writes this on LinkedIn :
We scaled Fish Audio to $13M ARR in 14 months with a very lean team. The entire GTM engine was built by the same people shipping the product. For the first time, we're hiring dedicated roles to scale what we've built.
She writes this on LinkedIn :
https://www.linkedin.com/posts/rissacao_developer-relations-engineer-activity-7459674982504275968-r9LO
Quote
1. Hiring Founding Account Executive
We've been closing every enterprise deal ourselves. Zero churn so far. Now we need someone to take over the pipeline and build the sales playbook with us. https://lnkd.in/gcscjYmB
2. Hiring Founding Forward Engineer
Our customers need someone on the ground to ship our models into production. That includes call centers, healthcare, real-time voice agents etc. https://lnkd.in/gK3bHs-a
3. Hiring Product Marketing Specialist
We need someone who owns how Fish Audio shows up to the market. Launches, positioning, competitive moves, all of it from day one. https://lnkd.in/gnvVZM6S
4. Hiring Audio LLM Researcher
Our model outperforms every major competitor in blind tests. We’ve 7M+ users feeding real signal back into training. Now we need a researcher to build what comes next. https://lnkd.in/gJwFdn6W
5. Hiring Developer Relations Engineer
We need someone to become the technical face of Fish Audio for API developers, builders, and voice AI teams. Build demos, support developers hands-on, and create real community momentum around our platform. https://lnkd.in/gqUuvtgT
We're building the team that takes Fish Audio to the next stage. If any of these sound like you, apply or tag someone who should. Shijia Liao, Jiahua Liu
End Quote
Note:
If you are now wondering WHO is behind this Voice Cloning Ai... it's students who seem NOT to have any business sense or background in building a business. Let me tell you who they are :
- Jiahua Liu (student)
- Rissa Cao
- Shijia Liao (student)
Here's more word's from it's co founder, oh and there's a "ai generated" photo of them too which I found on the post.
Well.. let's continue on shall we ? I found this post with more in-depth details on Fish Audio :
https://www.allvoicelab.com/blog/fishing-for-voices-with-fish-audio-from-trump-ai-voice-to-lebron-ai-voice
which states the following :
From viral celebrity impressions to professional AI voice synthesis—unlock limitless creativity with on-demand Text-to-Speech and Voice Cloning.
AI voice technology has become a cultural phenomenon, dominating search trends with phrases like “Trump AI voice,” “LeBron AI voice,” and “Scarlett Johansson AI voice.” At the same time, niche communities are buzzing about the “trapaholics voice generator” and “Fish.audio AI Digimon” samples. Leading the charge is Fish Audio, a browser-based platform that combines Text-to-Speech, AI Voice Cloning, and interactive Dubbing tools into a playful sandbox for creators
Origins of Fish Audio and Shijia Liao’s Vision
Behind Fish Audio is Shijia Liao, whose passion for voice synthesis began in his teens with open-source projects like So-VITS-SVC and GPT-SoVITS. After graduating from the University of Maryland in 2023, he joined NVIDIA to scale Vision Foundation Models across large GPU clusters.
However, a deeply personal experience shifted his focus: when his long-term relationship ended in betrayal, Shijia realized the fragility of human connections and envisioned AI companions that would never let users down.
Driven by this insight, he departed NVIDIA and spent countless nights prototyping Fish Audio on a home rig of 4090 GPUs. His work centered on emotional Text-to-Speech and high-fidelity AI Voice Cloning—combining prosody modeling with semantic cues—so that each voice could convey warmth, humor, or suspense on demand. This dedication laid the foundation for Fish Audio’s intuitive platform, where creators can generate studio-quality AudioBooks, Video Translation, and interactive Dubbing experiences with simple, emotion-driven prompts.
Shijia Liao’s Leadership
As Fish Audio’s founder, Shijia Liao navigated his first true startup journey by embracing both the technical and human challenges of entrepreneurship. He learned early on that venture success depends on selecting partners who complement each other’s strengths and on building a team aligned around a common mission. Today, Fish Audio’s engineering team is defined by its passion, rapid execution, and unwavering commitment. Tasks move seamlessly through Linear tickets, with each engineer owning features end-to-end—from development to bug fixing—and weekly syncs ensure alignment without unnecessary meetings.
The team’s drive was put to the test at the end of 2023, when Shijia balanced model development with the search for a commercial co-founder. Tight budgets and limited compute resources tested everyone’s resilience, but the shared vision kept the team together. By early 2024, those collective efforts paid off: Fish Audio’s product matured, revenue growth accelerated, and the platform gained solid traction.
On the funding side, Fish Audio closed a financing round right before acceptance into the HF0 incubator. This back-to-back validation catapulted the company’s revenue to $4 million in mere months and lifted its valuation, paving the way for the next strategic raise post-S1 model launch.
Fish Audio’s Performance and Technical Roadmap
By early 2024, Fish Audio’s relentless iteration delivered tangible results: annual recurring revenue (ARR) soared to $4 million in just months, and monthly active users climbed from 50,000 to over 400,000. Key internal metrics—like a 60% reduction in re-generation rates and a 70% export rate—highlight the platform’s improved emotional Text-to-Speech and high-fidelity AI Voice Cloning. Looking ahead, the upcoming S1 model will enable fine-grained controls such as “add suspense here” or “blend warmth with excitement,” pushing the boundaries of on‑demand voice customization.
What Makes Fish Audio Stand Out?
Celebrity & Character Voices
With just a few seconds of input, users can produce convincing impersonations—whether it’s a "Kanye AI voice" snippet for a music parody or an anime-inspired track using customer-favorite Digimon models.
·Intuitive Interface
No steep learning curve; the web UI walks even first-time creators through generating AudioBooks, short-form video narration, or live Voice Changer effects seamlessly.
Community-Driven Samples
From politics to pop culture, Fish Audio’s library keeps pace with meme trends—ideal for social clips, TikTok parodies, and meme pages aiming for viral traction
Fish Audio in the Wild: Six Creative Scenarios
1. Political Satire
Bloggers seed the keyword "Trump AI voice" in tutorial posts, boosting their SEO while offering step-by-step demos on crafting satirical news clips.
2. Sports Highlights
Video editors searching "LeBron AI voice" find ready-made scripts to narrate NBA montages—driving fan engagement and search traffic.
3. Film Reviews
Channels covering Marvel blockbusters embed "Scarlett Johansson AI voice" for dynamic trailer recaps that resonate with cinephiles.
4. EDM Remixes
Producers rely on the "trapaholics voice generator" for vocal hooks, sampling AI-generated rap lines to create fresh beats.
5. Parody Music
Amateur artists experiment with “Kanye AI voice” demos, mixing familiar cadences into comedic song covers.
6. Nostalgic Fan Projects
Anime enthusiasts search “Fish.audio AI Digimon” to revive classic character voices for fan-made AMVs and podcasts.
Beyond the Meme: All Voice Lab’s Workbench Advantage
Fish Audio’s playful focus on quick celebrity packs and meme-driven samples has sparked creativity across the web. Yet, sustained content production demands a robust workbench—something All Voice Lab has engineered to excel where novelty ends and real-world workflows begin:
·Customizable Pipelines
Our workbench lets teams define end-to-end jobs, chaining steps like emotion annotation → voice rendering → format conversion, all with reusable templates.
Enterprise-Grade Licensing
Unlike ad-hoc novelty generators, All Voice Lab’s platform includes built‑in compliance checks and asset management, ensuring each AudioBook or dubbing project meets legal and quality standards.
Scalable API Orchestration
For studios and learning platforms, our backend can process thousands of Video Translation requests per hour without sacrificing audio fidelity.
By learning from Fish Audio’s rapid viral appeal, All Voice Lab has built a workbench that combines the best of both worlds—easy experimentation plus industrial-strength pipelines. Whether you need a quick “Trump AI voice” demo，a good Voice Changer, or a full series of localized AudioBooks, our integrated platform ensures creativity never stalls at the proof‑of‑concept stage
The Future of AI Voice Is Hybrid
Fish Audio’s playground approach has ignited mainstream curiosity. As the field matures, the boundary between novelty and production is starting to blur. Ideal platforms will combine quick-start celebrity packs with deep, license-compliant pipelines for long-term projects. Whether you’re chasing “trapaholics voice generator” virality or producing multi-episode AudioBooks, this hybrid model promises both viral buzz and sustainable quality.
All Voice Lab’s Vision for the Future
While Fish Audio captivates creators with its playful celebrity packs and meme-ready samples, All Voice Lab focuses on building a comprehensive voice workbench that bridges experimentation and production. Our vision is to democratize professional-grade voice technology—making advanced Text-to-Speech, AI Voice Cloning, Voice Changer, and Video Translation accessible within a single, scalable platform. We aim to empower podcasters, game developers, educators, and enterprises to craft emotionally rich audio experiences at scale, fostering deeper connections and seamless global content delivery.
Stay tuned for more insights and innovations as both Fish Audio and All Voice Lab continue shaping the future of AI voice technology.
Driven by this insight, he departed NVIDIA and spent countless nights prototyping Fish Audio on a home rig of 4090 GPUs. His work centered on emotional Text-to-Speech and high-fidelity AI Voice Cloning—combining prosody modeling with semantic cues—so that each voice could convey warmth, humor, or suspense on demand. This dedication laid the foundation for Fish Audio’s intuitive platform, where creators can generate studio-quality AudioBooks, Video Translation, and interactive Dubbing experiences with simple, emotion-driven prompts.
Shijia Liao’s Leadership
As Fish Audio’s founder, Shijia Liao navigated his first true startup journey by embracing both the technical and human challenges of entrepreneurship. He learned early on that venture success depends on selecting partners who complement each other’s strengths and on building a team aligned around a common mission. Today, Fish Audio’s engineering team is defined by its passion, rapid execution, and unwavering commitment. Tasks move seamlessly through Linear tickets, with each engineer owning features end-to-end—from development to bug fixing—and weekly syncs ensure alignment without unnecessary meetings.
The team’s drive was put to the test at the end of 2023, when Shijia balanced model development with the search for a commercial co-founder. Tight budgets and limited compute resources tested everyone’s resilience, but the shared vision kept the team together. By early 2024, those collective efforts paid off: Fish Audio’s product matured, revenue growth accelerated, and the platform gained solid traction.
On the funding side, Fish Audio closed a financing round right before acceptance into the HF0 incubator. This back-to-back validation catapulted the company’s revenue to $4 million in mere months and lifted its valuation, paving the way for the next strategic raise post-S1 model launch.
Fish Audio’s Performance and Technical Roadmap
By early 2024, Fish Audio’s relentless iteration delivered tangible results: annual recurring revenue (ARR) soared to $4 million in just months, and monthly active users climbed from 50,000 to over 400,000. Key internal metrics—like a 60% reduction in re-generation rates and a 70% export rate—highlight the platform’s improved emotional Text-to-Speech and high-fidelity AI Voice Cloning. Looking ahead, the upcoming S1 model will enable fine-grained controls such as “add suspense here” or “blend warmth with excitement,” pushing the boundaries of on‑demand voice customization.
What Makes Fish Audio Stand Out?
Celebrity & Character Voices
With just a few seconds of input, users can produce convincing impersonations—whether it’s a "Kanye AI voice" snippet for a music parody or an anime-inspired track using customer-favorite Digimon models.
·Intuitive Interface
No steep learning curve; the web UI walks even first-time creators through generating AudioBooks, short-form video narration, or live Voice Changer effects seamlessly.
Community-Driven Samples
From politics to pop culture, Fish Audio’s library keeps pace with meme trends—ideal for social clips, TikTok parodies, and meme pages aiming for viral traction
Fish Audio in the Wild: Six Creative Scenarios
1. Political Satire
Bloggers seed the keyword "Trump AI voice" in tutorial posts, boosting their SEO while offering step-by-step demos on crafting satirical news clips.
2. Sports Highlights
Video editors searching "LeBron AI voice" find ready-made scripts to narrate NBA montages—driving fan engagement and search traffic.
3. Film Reviews
Channels covering Marvel blockbusters embed "Scarlett Johansson AI voice" for dynamic trailer recaps that resonate with cinephiles.
4. EDM Remixes
Producers rely on the "trapaholics voice generator" for vocal hooks, sampling AI-generated rap lines to create fresh beats.
5. Parody Music
Amateur artists experiment with “Kanye AI voice” demos, mixing familiar cadences into comedic song covers.
6. Nostalgic Fan Projects
Anime enthusiasts search “Fish.audio AI Digimon” to revive classic character voices for fan-made AMVs and podcasts.
Beyond the Meme: All Voice Lab’s Workbench Advantage
Fish Audio’s playful focus on quick celebrity packs and meme-driven samples has sparked creativity across the web. Yet, sustained content production demands a robust workbench—something All Voice Lab has engineered to excel where novelty ends and real-world workflows begin:
·Customizable Pipelines
Our workbench lets teams define end-to-end jobs, chaining steps like emotion annotation → voice rendering → format conversion, all with reusable templates.
Enterprise-Grade Licensing
Unlike ad-hoc novelty generators, All Voice Lab’s platform includes built‑in compliance checks and asset management, ensuring each AudioBook or dubbing project meets legal and quality standards.
Scalable API Orchestration
For studios and learning platforms, our backend can process thousands of Video Translation requests per hour without sacrificing audio fidelity.
By learning from Fish Audio’s rapid viral appeal, All Voice Lab has built a workbench that combines the best of both worlds—easy experimentation plus industrial-strength pipelines. Whether you need a quick “Trump AI voice” demo，a good Voice Changer, or a full series of localized AudioBooks, our integrated platform ensures creativity never stalls at the proof‑of‑concept stage
The Future of AI Voice Is Hybrid
Fish Audio’s playground approach has ignited mainstream curiosity. As the field matures, the boundary between novelty and production is starting to blur. Ideal platforms will combine quick-start celebrity packs with deep, license-compliant pipelines for long-term projects. Whether you’re chasing “trapaholics voice generator” virality or producing multi-episode AudioBooks, this hybrid model promises both viral buzz and sustainable quality.
All Voice Lab’s Vision for the Future
While Fish Audio captivates creators with its playful celebrity packs and meme-ready samples, All Voice Lab focuses on building a comprehensive voice workbench that bridges experimentation and production. Our vision is to democratize professional-grade voice technology—making advanced Text-to-Speech, AI Voice Cloning, Voice Changer, and Video Translation accessible within a single, scalable platform. We aim to empower podcasters, game developers, educators, and enterprises to craft emotionally rich audio experiences at scale, fostering deeper connections and seamless global content delivery.
Stay tuned for more insights and innovations as both Fish Audio and All Voice Lab continue shaping the future of AI voice technology.
---
OMFG !