While AI has made video creation widely accessible, many tools still overlook the importance of audio. Mirelo is developing AI designed to generate soundtracks that align with on-screen action. Earlier this year, the Berlin startup launched Mirelo SFX v1.5, an AI model that analyzes videos to add synchronized sound effects. This innovation captured the interest of venture capitalists preparing for a generative AI transformation in gaming. The fresh funding will strengthen Mirelo’s competitive position in this growing niche.

During its earlier stealth phase, when resources were limited, several larger players released their own video-to-sound effects models. These included Sony, Tencent, China’s Kuaishou-owned Kling AI, and ElevenLabs, which is backed by a16z. Although Mirelo already distinguishes itself through a more specialized focus, sustaining an edge against these models long-term will require expanding its team. New hires will accelerate research and development while advancing product strategy and market entry.

The startup has made its models available on platforms like Fal.ai and Replicate, anticipating that API usage will form the core of its near-term revenue, according to Simon-Gabriel. Simultaneously, it is investing in Mirelo Studio, a dedicated workspace for creators that may eventually support full professional workflows.

As Mirelo scales, the company and its investors are mindful of the training data concerns that have challenged other generative AI firms. Georgia Stevenson, who led the investment from Index, noted that Mirelo’s models are built on public and licensed sound libraries, and the company is establishing revenue-sharing agreements that honor artists’ rights. This tension is inherent to generative AI, but Mirelo isn’t replacing musicians and sound designers—at least not yet.

Operating on a freemium model with a recommended creator plan priced at €20 per month (about $23.50), Mirelo primarily serves amateurs and prosumers looking to add sound to AI-generated videos. Simon-Gabriel emphasized that audio is essential to unlocking the full potential of video. “George Lucas said sound is 50% of the movie-going experience. That’s not an overstatement,” he remarked. “If anything, it’s an understatement. The same visuals can evoke completely different atmospheres simply by changing the sound and music.”

Both Simon-Gabriel and his co-founder Florian Wenzel are AI researchers and musicians, and AI music generation remains on the company’s roadmap. However, current demand leans more toward sound effects, partly because this area has seen less research compared to other AI domains. “It’s easier to build a real moat here and then capitalize on it,” Simon-Gabriel observed.

This focus could prove advantageous. While Simon-Gabriel did not disclose the new valuation, he confirmed it has risen “very significantly” compared to the earlier undisclosed pre-seed round. That initial round was led by Berlin’s Atlantic Labs, which also joined the latest funding, bringing total investment to $44 million and helping close previous resource gaps.

Mirelo also benefits from angel investors who lend technical credibility and open doors, including Mistral CEO Arthur Mensch, Hugging Face chief science officer Thomas Wolf, and Fal.ai co-founder Burkay Gur. Still, the team recognizes that AI-generated videos may not remain silent for long—for example, Gemini’s video generator now includes soundtracks powered by DeepMind’s Veo 3.1 model.

Nevertheless, Simon-Gabriel feels validated by the growing attention to audio. “Now, suddenly, people realize, ‘Oh, maybe we should add sound.’ But of course you should. It’s like silent movies versus talkies—it makes quite a difference,” he said.