With a focus on synced sound effects and creator tools, the Berlin startup is tackling generative video’s biggest blind spot — and building a business around unmuting AI content.
The Silent Gap in AI Video
As AI-generated video tools evolve at a breakneck pace, one critical component is often left behind: sound.
While text-to-video models can now animate lifelike scenes, many lack real-time audio synchronization — a problem Berlin-based Mirelo is aiming to solve.
The startup’s pitch is simple but powerful: AI videos need sound to be complete. Their tool, Mirelo SFX v1.5, interprets video content to generate and sync sound effects (SFX), bringing AI visuals to life in a way that feels cinematic — and real.
“Sound is 50% of the experience,” says co-founder and CEO CJ Simon-Gabriel, echoing George Lucas. “It’s not an overstatement. If anything, it’s an understatement.”
A Big Bet on AI Audio
Now, investors are tuning in. Index Ventures and Andreessen Horowitz (a16z) have led a $41 million seed round in Mirelo, TechCrunch exclusively reports, bringing its total funding to $44 million.
The raise puts Mirelo in direct competition with larger, well-resourced players like Sony, Tencent, and ElevenLabs, which have also debuted video-to-audio models.
But Mirelo’s narrow focus on SFX (not dialogue or music) may give it an edge. While many competitors are chasing AI music or full video suites, Mirelo is targeting the unglamorous but essential layer of sound design: footsteps, doors, rustling leaves, explosions — the subtle cues that shape a viewer’s emotional experience.
AI With a Moat: Why Sound Effects Matter
According to Simon-Gabriel, SFX is less saturated in the AI space compared to music or voice generation. That’s both a strategic advantage and a technical one:
- Fewer competitors mean a clearer path to differentiation.
- The problem space allows Mirelo to build a defensible moat, using both proprietary models and custom training data.
That’s already paying off: Mirelo’s tools are publicly available on Fal.ai and Replicate, with API access driving early developer and creator adoption.
Scaling with Sound: From Tools to Studio
Mirelo’s 10-person team is set to double or triple in 2025, with new hires supporting:
- R&D for AI model refinement
- Go-to-market and product growth
- Expansion of its creator-focused app, Mirelo Studio
Studio aims to provide a full workspace for video creators, making it easy to add, customize, and sync audio directly within their production pipeline — no technical expertise needed.
Mirelo’s current freemium model includes a €20/month creator tier, targeting prosumers and indie creators who need scalable sound solutions but don’t have Hollywood budgets.
Rights, Royalties, and the Ethics of Training Data
As the generative AI space faces growing scrutiny over copyrighted training data, Mirelo is preemptively working to stay clean.
According to Index Ventures’ Georgia Stevenson, the company has:
- Used public and licensed sound libraries for training
- Signed revenue-sharing deals to compensate artists
- Avoided scraping or unlicensed data in model development
That could be a long-term differentiator — especially if regulatory pressure on AI model transparency increases.
Looking Ahead: From SFX to Music?
While sound effects are Mirelo’s current focus, Simon-Gabriel — both an AI researcher and musician — says AI-generated music is on the roadmap.
But for now, SFX is where the market is pulling the company. It’s also where Mirelo believes it can deliver professional-grade results at scale, far earlier than fully automated soundtracks or dialogue.
“It’s easier to build a real moat here, and then to capitalize on it,” Simon-Gabriel noted.
Competitive Winds Are Shifting
Mirelo may have been early, but others are catching on.
- DeepMind’s Veo 3.1, now part of Gemini, integrates video-to-audio.
- Kling AI and other Chinese firms are racing to build end-to-end video tools, including audio.
Yet Simon-Gabriel sees this less as a threat and more as validation:
“It’s a bit like silent movies versus talkies, right? It does make quite a difference.”
With key angel support from Mistral’s CEO, Hugging Face’s CSO, and Fal.ai’s co-founder, Mirelo has the technical credibility and funding to make that difference — and possibly define the sound layer of AI video as it emerges.
