Mirelo is an AI company based out of Berlin which uses foundation models to create sound effects in sync with video content. Mirelo has filled a void in AI development by allowing video creators to create matching audio in seconds as opposed to spending hours searching through stock libraries and manually syncing.
Founded by CJ Simon-Gabriel (CEO) and Florian Wenzel (CTO) who are AI researchers from Amazon Web Services Labs. They started Mirelo to build foundation models for generating synchronized sound effects for videos.
Initial funding round closed by Atlantic VC, while working in stealth mode with limited resources.
Major seed round secured with Index Ventures and Andreessen Horowitz as lead investors along with participation from Atlantic.vc and TriplePoint Capital, resulting in total of $44 million raised.
Published foundation models on the Fal.ai and Replicate platforms to start generating API-driven revenue.
Announced the Mirelo Studio workspace for creators, announced roadmap for AI music generation and announced they will be hiring additional team members to grow their team from 10 to 20-30+ employees.
Specialized foundation models for sound effect generation with advanced tokenization, data curation, and conditioning techniques that represent breakthroughs in generative audio synthesis.
Based on official announcements and investor statements; detailed technical architecture not publicly disclosed
| Service | Cost | Details | Source |
|---|---|---|---|
| Free | €0 per month | 2,400 credits/month, latest Mirelo audio models, 3 projects, projects delete after 1 week, watermark on exports | F6S pricing page |
| Starter | €5 per month | 6,000 credits/month, latest Mirelo audio models, 6 projects, no credit top-ups | F6S pricing page |
| Creator | €20 per month | 22,000 credits/month, latest Mirelo audio models, 20 projects, additional credits €0.90/1000 | F6S pricing page |
| Studio | €99 per month | 120,000 credits/month, latest Mirelo audio models, unlimited projects, additional credits €0.90/1000 | F6S pricing page |
| Scale | €299 per month | 400,000 credits/month, latest Mirelo audio models, unlimited projects, additional credits €0.80/1000 | F6S pricing page |
| Enterprise | €999 per month or custom | 1,440,000 credits/month (standard), unlimited projects, additional credits €0.75/1000, volume discounts and custom terms available | F6S pricing page |
| Feature | Mirelo AI | MusicCreator AI | Soundraw | Lalals |
|---|---|---|---|---|
| Core Functionality | AI Audio Generation (music & SFX) | AI Music, Lyrics, Vocals, Stem Splitter | AI Music Generation | AI Music Generation |
| Pricing (starting paid) | €5/mo | $10.49/mo | $10/mo | $11.99/mo |
| Free Tier | Yes (2,400 credits) | Yes (2 songs/mo) | No | No |
| Enterprise Features | Custom terms, volume discounts | |||
| API Availability | ||||
| Credit-based Usage | Yes (credits/month) | Songs & generations | Generation credits | |
| Support Options | — | Email (24/7 on higher tiers) | ||
| Watermark on Free | Yes |
Mirelo is an audio-generating service that focuses on providing both music and SFX through a credit-based system that allows users to generate their own audio through a free plan that includes watermarks, whereas MusicCreator's plan also gives users more tools like lyrics and vocal removal, however, it has limits on how much of a song you can produce before reaching your limit. Mirelo is currently the cheaper option at €5 compared to MusicCreator's price of $10.49 per month, which is based on credits rather than songs.
Mirelo would best fit the needs of those who need flexible audio or SFX creation, and MusicCreator would be the best choice for creating complete song production workflows.
Both are AI-music generators, however, Soundraw does require payment for its services ($10 per month, with no free version), while Mirelo provides a very generous free version that is limited by the number of credits the user has available each month. In addition, Mirelo will allow users to scale their subscription to meet the needs of their business as they grow.
Mirelo is best suited for those who want to test or experiment with different types of audio/SFX creation, while Soundraw is best for those who are serious about being a creator.
The two companies have similar entry-level prices (Mirelo is €5 compared to LalalSound's $11.99 per month), however, Mirelo offers a variety of tiered levels of pricing, up to the level of enterprise, with custom pricing options for larger organizations, while LalalSound appears to be geared towards music creation only and does not appear to have the ability to scale.
Mirelo is best suited for large groups of people who need to create large amounts of audio, while LalalSound is best suited for individuals who want to create music.
A silent video can be uploaded to Mirelo's application or API which then uses an artificial intelligence to analyze the video content and produce synchronized sound effects and music in seconds. The editor allows for customizable timing, ambience, and effects. The final product is ready to be used in film/TV productions with the audio perfectly matching the video action.
There are no published prices for Mirelo. Sales contact information is required to obtain a quote. Mirelo also provides access to its API, which can be tested by registering on pay per use platforms such as Replicate.com and Fal.ai. Enterprise users may have additional options for model fine tuning.
Mirelo focuses on creating sound effects that synchronize with video motion, context, and timing. The resulting SFX sound effects can be similar to those found in a major Hollywood movie. Mirelo is focused solely on creating sound effects that match the action of the video footage. In contrast, most other music generators create complete music tracks that do not provide a close match to the video action.
Mirelo does not mention any type of security certification on their website. Mirelo processes video data for sound effect creation purposes only. When using the Mirelo API, videos should be silent and less than 10 seconds long. Reviewing the terms of service provided by Mirelo for production use will require you to understand the security measures they have implemented.
Yes, the Mirelo API can be accessed via a REST API call to api.mirelo.ai/video-to-sfx with an API key for authentication. The Mirelo API is also accessible on Replicate.com and Fal.ai for easy testing. Model fine tuning is available for custom sound asset use.
Video lengths longer than 10 seconds will be truncated to 10 seconds. It is recommended to upload silent videos to the Mirelo API. Users can also add text prompts to guide the selection of sound effects, for example "dog barking", and select from multiple output variations of 2-4 samples.
There is no publicly listed free version of Mirelo. Testing of the Mirelo API is available at the Replicate.com model page with pay per run pricing. Obtaining a full API access key from mirelo.ai is necessary to test the Mirelo API for all features.
Mirelo offers model fine tuning with your company's existing sound assets for enterprise level clients. For further details, please contact sales for custom training and integration options. The Mirelo API accepts text prompts to provide a stylistic guideline when selecting sound effects.
Mirelo provides a breakthrough solution for video-to-SFX generation with accurate synchronization and cinematic-quality results and solves an essential gap in the AI video production workflow. The $41 million funding received from leading venture capital firms, combined with its lightweight architecture (50 times less computing power), positions it well within the developing market for procedural audio. At early-stage status with very little publicly available documentation, but clearly offers large-scale enterprise potential.
Mirelo is best suited for video creators and platforms that are working to solve the "AI silent video" problem by generating precise, context-sensitive sound effects.
Mirelo is a developer of artificial intelligence for video-to-sfx/music generation with precise synchronization. Their flagship model, SFX v1.5 uses motion/context analysis to provide cinematic-quality audio. With $41 million in funding from Index Ventures and a16z validating their market-fit, they have made their API accessible through api.mirelo.ai and platforms such as Replicate/Fal.ai. They focus primarily on the enterprise fine-tuning and app integration aspects of their technology.
Fair - strong signal from official site and TechCrunch funding news. Limited transparency on pricing/security/SLA; no dedicated docs/API portal or status page found publicly.
Received $41 million from Index Ventures and a16z (Dec 25) to grow their AI video to audio capabilities. A Berlin-based startup that is using artificial intelligence to address the "silent crisis" in AI video production with procedural SFX generation.
Mirelo SFX v1.5 is available on both Replicate.com and Fal.ai for ease of use through their API. Silent video input is supported up to 10 seconds. Text prompts are also supported as well as multiple variants. All output from Mirelo SFX is ready for production and artifact free.
In 2025 Dr. Florian Wenzel, founder and CEO, gave a presentation at the dida conference about video aware music and SFX generation using foundation models. The focus was on how multimodal AI could be used to understand the pacing and mood of a video.
Mirelo SFX has been trained on a combination of publicly available and purchased sound libraries in addition to partnering with artists in an effort to create a revenue sharing relationship. This allows them to address the ethical issues associated with using copyrighted material while also building a proprietary video-SFX alignment.
Mirelo SFX requires approximately 50 times less compute than other typical LLMs, and it can generate video-SFX alignment at a rate that is 1.7 times faster than real time. Mirelo SFX provides contextual understanding and this allows it to distinguish between gentle taps, forceful impacts, and scene transitions.
Generate synchronized sound effects directly from video content analysis.
Use text prompts such as metal clanging or birds chirping to customize sound generation.
Time your audio effects so they precisely match the timing of the visual action and motion.
Produce 2-4 variants of audio per video to allow for selection and creativity.
Scene analysis, motion intensity, and interaction of objects by AI
Editing after AI-generated media as a fully creative process.
Background sounds and atmosphere generation.
Video-to-audio creation integrated into customized apps.
Synchronized SFX to TikTok, YouTube Shorts, Instagram Reels.
Automated Foley & Background SFX Generation.
Silent AI generated videos transformed into engaging media.
Procedural audio effects that match game play footage.
Rapid turnaround SFX for video campaign projects.
Automation of secondary sound effects
