Voicemod bags $14.5M to bring real-time AI voices to everyone

Voicemod just raised $14.5M to give everyone an AI-powered voice makeover.

The startup behind the popular voice-changing desktop app announced a $14.5 million expansion round led by Leadwind (Kfund), with backing from former Discord CMO Eros Resmini's Minifund and Bitkraft Ventures. Founded in 2014, Voicemod already counts over 40 million desktop downloads and 3.3 million monthly active users.

Now, it is pivoting hard from traditional digital signal processing into real-time generative AI.

Putting on an audio avatar

For nearly a decade, Voicemod mostly made digital audio effects: silly soundboards, robotic distortions, and audio memes for gamers on Discord. That was digital signal processing—applying filters to your actual voice.

Generative AI flips that entirely. Instead of tweaking your vocal cords, Voicemod is building tools that let you wear a completely synthetic voice in real time, like an audio Snapchat lens.

That includes what CEO Jamie Bosch calls sing-to-sing voice conversion. You could sing off-key into your mic, and the engine transforms your output live into the voice of Freddie Mercury, complete with proper pitch.

The push comes after Voicemod acquired audio startup Voctro Labs last year, the team behind the viral Holly Herndon AI voice clone. Voicemod already shipped a text-to-song feature in December and is combining Voctro's synthesis models with its own real-time DSP to roll out live speech-to-speech engines. Beyond its consumer app, the company distributes its tech to games, apps, and hardware makers via SDKs and APIs.

Why it matters

Voice is becoming just another software layer you can swap out at will.

For creators and casual users, the barrier to audio production is evaporating. You won't need professional gear or classical vocal training to drop a polished track, narrate a YouTube short, or roleplay a character in an MMO. Bosch even envisions everyday audio makeup for cleaning up how you sound on work Zoom calls.

Strategically, Voicemod has a massive head start over pure-play AI research labs. With 3.3 million monthly active users already in its Discord-friendly gamer sandbox, it does not need to hunt for distribution. It just has to layer real-time AI into a workflow its audience already uses every night.

We think the immediate risk here lies in security and identity. As real-time voice swapping goes mainstream, voiceprint-based banking authentication and phone security are effectively broken. In a world where anyone can wear anyone else's voice on a live call, proof of human identity is about to get much messier.