HeadsUpAI

fal Launches Mirelo SFX 1.6 for Synchronized AI Video Sound Design

fal, a generative media infrastructure platform for serverless inference, launched Mirelo SFX 1.6. This suite specializes in Foley work—creating environmental sounds—and includes tools for video-to-audio synchronization, audio inpainting, and extending existing clips with seamless tails. Each capability is accessible via a dedicated API endpoint for automated workflows.

While frontier video models produce high-fidelity visuals, they typically lack native audio. This release follows the launch of fal's genmedia CLI, which brought media generation to the terminal. By adding inpainting and looping, Mirelo 1.6 moves beyond basic text-to-audio to provide the granular control required for professional video editing.

You can use the Ambiancer tool to create looping soundscapes or the Inpainter to replace specific audio segments. These capabilities are available now via the fal API and playground, with inference priced at $0.01 per second. This modular approach allows you to add sound design to any video without regenerating the visuals.

View the full update on fal.ai
fal
fal
@fal
X

🚨 Mirelo SFX 1.6 is now live on fal! 🎬 Video → synced sound effects (up to 60s) ➕ Extend any sound with seamless tails 🌿 Ambiancer: natural looping soundscapes ✂️ Inpainter: erase & replace unwanted audio https://t.co/jM4FSjQvsv

4retweets31likes
View on X

Still wondering? A few quick answers below.

Mirelo SFX 1.6 is a specialized AI audio model suite available on the fal platform designed for high-fidelity sound effects and ambient audio. Unlike general music generators, it focuses on Foley work, which involves creating synchronized, environmental, and situational sounds for visual media like AI-generated videos that are typically silent.

The video-to-audio capability allows users to upload a video file and generate synchronized sound effects that match the visual action. The model can process and add audio to video clips up to 60 seconds long, providing a dedicated sound design layer that aligns with the timing and context of the original footage.

Audio inpainting is a functional editing tool that allows users to erase specific, unwanted parts of an existing audio file and replace them with new generated content. This capability transforms the system from a simple generation tool into a post-production editor for fixing or modifying sound effects and ambient tracks.

Accessing the Mirelo SFX 1.6 models on the fal platform costs $0.01 per second of generated audio. This pricing applies across the different model variants, including text-to-audio, audio inpainting, audio extension, and the video-to-video synchronization tool, and is available through both the fal API and the web-based playground.

Yes, the suite includes a specific tool called Ambiancer designed to create natural, looping soundscapes. This feature allows users to turn any generated sound effect or text prompt into a seamless loop, which is particularly useful for creating consistent background environments and atmospheric audio for long-form media.

Share this update